mirror_edk2.git
4 years agoMdeModulePkg/UsbBusDxe: Add UsbControlTransfer() error check
Michael D Kinney [Tue, 14 Nov 2017 23:41:08 +0000 (15:41 -0800)]
MdeModulePkg/UsbBusDxe: Add UsbControlTransfer() error check

https://bugzilla.tianocore.org/show_bug.cgi?id=767

The USB I/O Protocol function ControlTransfer() has a DataLength
parameter that specifies the size of the Data buffer.  The
UsbBusDxe module implements the USB I/O Protocol using the
services of the USB2 Host Controller Protocol.  The DataLength
parameter in the USB2 Host Controller Protocol ControlTransfer()
service is an IN OUT parameter so the number of bytes actually
transferred is returned.  Since the USB I/O Protocol
ControlTransfer() service can not return the number of bytes
actually transferred, the only option if the number of bytes
actually transferred is less than the number of bytes requested
is to return EFI_DEVICE_ERROR.

The change fixes an issue with a USB mass storage device that
responds with 0 bytes to the Get MAX LUN command.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoBaseTools: Fix a bug for single module build with GenC/GenMake option
Yonghong Zhu [Fri, 17 Nov 2017 03:29:19 +0000 (11:29 +0800)]
BaseTools: Fix a bug for single module build with GenC/GenMake option

when build a single module with GenC/GenMake option, currently it will
direct return after create Autogen code files, then it cause MaList is
empty, which cause an incorrect error message is reported.

Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yonghong Zhu <yonghong.zhu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoBaseTools: Guid.xref contain information from FILE statements in FDF
Yonghong Zhu [Wed, 15 Nov 2017 16:06:45 +0000 (00:06 +0800)]
BaseTools: Guid.xref contain information from FILE statements in FDF

Update Guid.xref to contain information from FILE statements in FDF
file.

Fixes: https://bugzilla.tianocore.org/show_bug.cgi?id=778
Cc: Liming Gao <liming.gao@intel.com>
Cc: Dmitry Antipov <dmanti@microsoft.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yonghong Zhu <yonghong.zhu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg/Core: Fix build error with old Visual Studio
Jian J Wang [Mon, 20 Nov 2017 02:59:41 +0000 (10:59 +0800)]
MdeModulePkg/Core: Fix build error with old Visual Studio

The build error is introduced by following check in:
  2930ef9809976ce693d1d377851344c3b06bd926
  235a4490c8ce8b6dbac49e6ae3559cb73d6bf620

The Visual Studio older than 2015 doesn't support constant integer
in binary format (0bxxx). This patch changes them to BIT macro to
fix it. This patch also cleans up coding style about unmatched
comment for return value.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Bi Dandan <dandan.bi@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg/PiSmmCore: Fix an assert issue in pool free
Jian J Wang [Mon, 20 Nov 2017 01:00:27 +0000 (09:00 +0800)]
MdeModulePkg/PiSmmCore: Fix an assert issue in pool free

Following check-in caused this issue:
  2930ef9809976ce693d1d377851344c3b06bd926

This issue is caused by assigning incorrect value to PoolHdr.Size
if the requested size is more than MAX_POOL_SIZE, because the SMM
core will actually allocate page memory instead of pool memory in
this situation.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Wu Hao <hao.a.wu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Wu Hao <hao.a.wu@intel.com>
Tested-by: Wu Hao <hao.a.wu@intel.com>
4 years agoBaseTools: Fix the bug to collect source files per build rule family
Yonghong Zhu [Mon, 13 Nov 2017 07:24:17 +0000 (15:24 +0800)]
BaseTools: Fix the bug to collect source files per build rule family

when collect source files list we should also consider build rule
family. BuildRuleFamily may be set to the different one. It will
impact BuildRule and source files in INF file.

Fixes: https://bugzilla.tianocore.org/show_bug.cgi?id=780
Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yonghong Zhu <yonghong.zhu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg/UdfDxe: Avoid possible loss track of allocated buffer
Hao Wu [Tue, 14 Nov 2017 03:33:10 +0000 (11:33 +0800)]
MdeModulePkg/UdfDxe: Avoid possible loss track of allocated buffer

In function FindFileEntry():

Instead of using the function parameter 'FileEntry', use a local
variable to store the buffer allocated for disk read operation.

For the below calling stack:
UdfOpenVolume() -> FindRootDirectory() -> FindFileEntry()

In FindFileEntry(), the call to 'DiskIo->ReadDisk()' is possible (e.g.
media change for a CD/DVD ROM device) to trigger a re-install of the
BlockIO(2) protocol which will further lead to a call of the BindingStop()
& BingdingStart() of the UdfDxe driver.

Meanwhile, for the above listed calling stack, the '**FileEntry'
parameter passed into FindFileEntry() is '&PrivFsData->Root'. 'PrivFsData'
is a driver-managed private data, it will be freed in BindingStop() and
re-allocate in BingdingStart().

In such case, if '*FileEntry' is used to store the allocated buffer, the
information will be lost if 'DiskIo->ReadDisk()' triggers a re-install of
the BlockIO(2) protocol. The subsequent call of the FreePool API:

FreePool (*FileEntry);

will cause issues.

This commit uses a local variable to store the allocated buffer.

Cc: Ruiyu Ni <ruiyu.ni@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: Paulo Alcantara <pcacjr@zytor.com>
4 years agoMdeModulePkg/PartitionDxe: Merge the discovery of ElTorito into UDF
Hao Wu [Fri, 13 Oct 2017 07:14:51 +0000 (15:14 +0800)]
MdeModulePkg/PartitionDxe: Merge the discovery of ElTorito into UDF

In order to create all of the children (El Torito standard and UDF) for
a CD/DVD media in an entry of the PartitionDriverBindingStart(), this
commit merges the discovery of the El Torito feature
(PartitionInstallElToritoChildHandles) into function
PartitionInstallUdfChildHandles.

Cc: Ruiyu Ni <ruiyu.ni@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: Paulo Alcantara <pcacjr@zytor.com>
4 years agoArmPlatformPkg/ArmPlatformLibNull: remove bogus PCD dependencies
Ard Biesheuvel [Fri, 17 Nov 2017 15:40:38 +0000 (15:40 +0000)]
ArmPlatformPkg/ArmPlatformLibNull: remove bogus PCD dependencies

Remove dependencies on gArmTokenSpaceGuid.PcdSystemMemoryBase and
gArmTokenSpaceGuid.PcdSystemMemorySize, the presence of which in
a [FixedPcd] section makes this module unusable for ArmVirtQemu.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoMdeModulePkg/UsbMassStorageDxe: Enhance Request Sense Handling
Michael D Kinney [Tue, 14 Nov 2017 23:56:30 +0000 (15:56 -0800)]
MdeModulePkg/UsbMassStorageDxe: Enhance Request Sense Handling

https://bugzilla.tianocore.org/show_bug.cgi?id=782

Update the Request Sense check for the Request Sense Key of
USB_BOOT_SENSE_UNIT_ATTENTION.  For this Sense Key, the
Additional Sense Key to EFI_STATUS mappings are:

USB_BOOT_ASC_MEDIA_CHANGE -> EFI_MEDIA_CHANGE
USB_BOOT_ASC_NOT_READY    -> EFI_NOT_READY
USB_BOOT_ASC_NO_MEDIA     -> EFI_NOT_READY
All others                -> EFI_DEVICE_ERROR

A USB flash drive is returning Request Sense Key of
USB_BOOT_SENSE_UNIT_ATTENTION and an Additional Sense Key of
USB_BOOT_ASC_NO_MEDIA for a few seconds before returning an
Additional Sense Key of USB_BOOT_ASC_MEDIA_CHANGE.

The current logic treats this initial Request Sense info as an
error and reties the command 5 times before failing completely.

With this change the USB Flash Drive works correctly.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoOvmfPkg: save on I/O port accesses when the debug port is not in use
Paolo Bonzini [Thu, 16 Nov 2017 20:31:00 +0000 (21:31 +0100)]
OvmfPkg: save on I/O port accesses when the debug port is not in use

When SEV is enabled, every debug message printed by OVMF to the
QEMU debug port traps from the guest to QEMU character by character
because "REP OUTSB" cannot be used by IoWriteFifo8.  Furthermore,
when OVMF is built with the DEBUG_VERBOSE bit (value 0x00400000)
enabled in "gEfiMdePkgTokenSpaceGuid.PcdDebugPrintErrorLevel", then the
OvmfPkg/IoMmuDxe driver, and the OvmfPkg/Library/BaseMemEncryptSevLib
library instance that is built into it, produce a huge amount of
log messages.  Therefore, in SEV guests, the boot time impact is huge
(about 45 seconds _additional_ time spent writing to the debug port).

While these messages are very useful for analyzing guest behavior,
most of the time the user won't be capturing the OVMF debug log.
In fact libvirt does not provide a method for configuring log capture;
users that wish to do this (or are instructed to do this) have to resort
to <qemu:arg>.

The debug console device provides a handy detection mechanism; when read,
it returns 0xE9 (which is very much unlike the 0xFF that is returned by
an unused port).  Use it to skip the possibly expensive OUT instructions
when the debug I/O port isn't plugged anywhere.

For SEC, the debug port has to be read before each full message.
However:

- if the debug port is available, then reading one byte before writing
a full message isn't tragic, especially because SEC doesn't print many
messages

- if the debug port is not available, then reading one byte instead of
writing a full message is still a win.

Contributed-under: TianoCore Contribution Agreement 1.0
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen (Intel address) <jordan.l.justen@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
4 years agoOvmfPkg: create a separate PlatformDebugLibIoPort instance for SEC
Paolo Bonzini [Thu, 16 Nov 2017 20:30:59 +0000 (21:30 +0100)]
OvmfPkg: create a separate PlatformDebugLibIoPort instance for SEC

The next patch will want to add a global variable to
PlatformDebugLibIoPort, but this is not suitable for the SEC
phase, because SEC runs from read-only flash.  The solution is
to have two library instances, one for SEC and another
for all other firmware phases.  This patch adds the "plumbing"
for the SEC library instance, separating the INF files and
moving the constructor to a separate C source file.

Contributed-under: TianoCore Contribution Agreement 1.1
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen (Intel address) <jordan.l.justen@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
4 years agoOvmfPkg: make PlatformDebugLibIoPort a proper BASE library
Paolo Bonzini [Thu, 16 Nov 2017 20:30:58 +0000 (21:30 +0100)]
OvmfPkg: make PlatformDebugLibIoPort a proper BASE library

Remove Uefi.h, which includes UefiSpec.h, and change the
return value to match the RETURN_STATUS type.

Contributed-under: TianoCore Contribution Agreement 1.1
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen (Intel address) <jordan.l.justen@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
4 years agoOvmfPkg: restore temporary SEC/PEI RAM size to 64KB
Laszlo Ersek [Tue, 24 Oct 2017 15:21:25 +0000 (17:21 +0200)]
OvmfPkg: restore temporary SEC/PEI RAM size to 64KB

(1) In the PEI phase, the PCD database is maintained in a GUID HOB. In
    OVMF, we load the PCD PEIM before any other PEIMs (using APRIORI PEI),
    so that all other PEIMs can use dynamic PCDs. Consequently,

    - the PCD GUID HOB is initially allocated from the temporary SEC/PEI
      heap,

    - whenever we introduce a dynamic PCD to a PEIM built into OVMF such
      that the PCD is new to OVMF's whole PEI phase, the PCD GUID HOB (and
      its temporary heap footprint) grow.

    I've noticed that, if we add just one more dynamic PCD to the PEI
    phase, then in the X64 build,

    - we get very close to the half of the temporary heap (i.e., 8192
      bytes),

    - obscure PEI phase hangs or DXE core initialization failures
      (ASSERTs) occur. The symptoms vary between the FD_SIZE_2MB and
      FD_SIZE_4MB builds of X64 OVMF.

(2) I've found that commit

      2bbd7e2fbd4b ("UefiCpuPkg/MtrrLib: Update algorithm to calculate
                    optimal settings", 2017-09-27)

    introduced a large (16KB) stack allocation:

> The patch changes existing MtrrSetMemoryAttributeInMtrrSettings() and
> MtrrSetMemoryAttribute() to use the 4-page stack buffer for calculation.
> ...
> +#define SCRATCH_BUFFER_SIZE           (4 * SIZE_4KB)
> ...
> @@ -2207,17 +2462,66 @@ MtrrSetMemoryAttributeInMtrrSettings (
> ...
> +  UINT8                      Scratch[SCRATCH_BUFFER_SIZE];

(3) OVMF's temp SEC/PEI RAM size has been 32KB ever since commit

      7cb6b0e06809 ("OvmfPkg: Move SEC/PEI Temporary RAM from 0x70000 to
                    0x810000", 2014-01-21)

    Of that, the upper 16KB half is stack (growing down), and the lower
    16KB half is heap.

    Thus, OvmfPkg/PlatformPei's calls to "UefiCpuPkg/Library/MtrrLib", in
    QemuInitializeRam(), cause the Scratch array to overflow the entire
    stack (heading towards lower addresses), and corrupt the heap below
    the stack. It turns out that the total stack demand is about 24KB, so
    the overflow is able to corrupt the upper 8KB of the heap. If that
    part of the heap is actually used (for example because we grow the PCD
    GUID HOB sufficiently), mayhem ensues.

(4) Right after commit 7cb6b0e06809 (see above), there would be no room
    left above the 32KB temp SEC/PEI RAM. However, given more recent
    commits

      45d870815156 ("OvmfPkg/PlatformPei: rebase and resize the permanent
                    PEI memory for S3", 2016-07-13)

      6b04cca4d697 ("OvmfPkg: remove PcdS3AcpiReservedMemoryBase,
                    PcdS3AcpiReservedMemorySize", 2016-07-12)

    we can now restore the temp SEC/PEI RAM size to the original
    (pre-7cb6b0e06809) 64KB. This will allow for a 32KB temp SEC/PEI
    stack, which accommodates the ~24KB demand mentioned in (3).

    (Prior patches in this series will let us monitor the stack usage in
    the future.)

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=747
Ref: http://mid.mail-archive.com/a49cc089-12ae-a887-a4d6-4dc509233a74@redhat.com
Ref: http://mid.mail-archive.com/03e369bb-77c4-0134-258f-bdae62cbc8c5@redhat.com
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
4 years agoOvmfPkg/Sec/X64: seed the temporary RAM with PcdInitValueInTempStack
Laszlo Ersek [Wed, 8 Nov 2017 19:32:30 +0000 (20:32 +0100)]
OvmfPkg/Sec/X64: seed the temporary RAM with PcdInitValueInTempStack

This allows the PEI core to report the maximum temporary SEC/PEI stack
usage on the DEBUG_INFO level, in the PeiCheckAndSwitchStack() function
[MdeModulePkg/Core/Pei/Dispatcher/Dispatcher.c]:

* Normal boot:

> Temp Stack : BaseAddress=0x814000 Length=0x4000
> Temp Heap  : BaseAddress=0x810000 Length=0x4000
> Total temporary memory:    32768 bytes.
>   temporary memory stack ever used:       5080 bytes. <----
>   temporary memory heap used for HobList: 8080 bytes.
>   temporary memory heap occupied by memory pages: 0 bytes.

* S3 resume (no SMM / PEI decompression)

> Temp Stack : BaseAddress=0x814000 Length=0x4000
> Temp Heap  : BaseAddress=0x810000 Length=0x4000
> Total temporary memory:    32768 bytes.
>   temporary memory stack ever used:       5048 bytes. <----
>   temporary memory heap used for HobList: 7112 bytes.
>   temporary memory heap occupied by memory pages: 0 bytes.

I unit-tested this change by transitorily adding an infinite loop right
after the "rep stosq", and dumping the guest's temp SEC/PEI RAM (32KB
currently) while the guest was stuck in the loop. The dump includes one
dword from before and after the temp SEC/PEI RAM:

> $ virsh qemu-monitor-command GUEST_NAME --hmp 'xp /8194wx 0x80FFFC'
>
000000000080fffc: 0x00000000 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5
000000000081000c: 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5
> ...
0000000000817fec: 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5
0000000000817ffc: 0x5aa55aa5 0x00000000

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=747
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
4 years agoOvmfPkg/Sec/Ia32: seed the temporary RAM with PcdInitValueInTempStack
Laszlo Ersek [Wed, 8 Nov 2017 18:43:21 +0000 (19:43 +0100)]
OvmfPkg/Sec/Ia32: seed the temporary RAM with PcdInitValueInTempStack

This allows the PEI core to report the maximum temporary SEC/PEI stack
usage on the DEBUG_INFO level, in the PeiCheckAndSwitchStack() function
[MdeModulePkg/Core/Pei/Dispatcher/Dispatcher.c]:

* Normal boot:

> Temp Stack : BaseAddress=0x814000 Length=0x4000
> Temp Heap  : BaseAddress=0x810000 Length=0x4000
> Total temporary memory:    32768 bytes.
>   temporary memory stack ever used:       3664 bytes. <----
>   temporary memory heap used for HobList: 5904 bytes.
>   temporary memory heap occupied by memory pages: 0 bytes.

* S3 resume (with PEI decompression / SMM):

> Temp Stack : BaseAddress=0x814000 Length=0x4000
> Temp Heap  : BaseAddress=0x810000 Length=0x4000
> Total temporary memory:    32768 bytes.
>   temporary memory stack ever used:       3428 bytes. <----
>   temporary memory heap used for HobList: 4816 bytes.
>   temporary memory heap occupied by memory pages: 0 bytes.

I unit-tested this change by transitorily adding an infinite loop right
after the "rep stosd", and dumping the guest's temp SEC/PEI RAM (32KB
currently) while the guest was stuck in the loop. The dump includes one
dword from before and after the temp SEC/PEI RAM:

> $ virsh qemu-monitor-command GUEST_NAME --hmp 'xp /8194wx 0x80FFFC'
>
000000000080fffc: 0x00000000 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5
000000000081000c: 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5
> ...
0000000000817fec: 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5 0x5aa55aa5
0000000000817ffc: 0x5aa55aa5 0x00000000

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=747
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
4 years agoUefiCpuPkg/ResetVector/Vtf0: document segment register setup
Laszlo Ersek [Wed, 15 Nov 2017 18:00:17 +0000 (19:00 +0100)]
UefiCpuPkg/ResetVector/Vtf0: document segment register setup

"Main.asm" calls TransitionFromReal16To32BitFlat (and does some other
things) before it jumps to the platform's SEC entry point.

TransitionFromReal16To32BitFlat enters big real mode, and sets the DS, ES,
FS, GS, and SS registers to offset ("selector") LINEAR_SEL in the GDT
(defined in "UefiCpuPkg/ResetVector/Vtf0/Ia16/Real16ToFlat32.asm"). The
GDT entry ("segment descriptor") at LINEAR_SEL defines a segment covering
the full 32-bit address space, meant for "read/write data".

Document this fact for all the affected segment registers, as output
parameters for TransitionFromReal16To32BitFlat, saying "Selector allowing
flat access to all addresses".

For 64-bit SEC, "Main.asm" calls Transition32FlatTo64Flat in addition,
between calling TransitionFromReal16To32BitFlat and jumping to the SEC
entry point. Transition32FlatTo64Flat enters long mode. In long mode,
segmentation is largely ignored:

- all segments are considered flat (covering the whole 64-bit address
  space),

- with the (possible) exception of FS and GS, whose bases can still be
  changed, albeit with new methods, not through the GDT. (Through the
  IA32_FS_BASE and IA32_GS_BASE Model Specific Registers, and/or the
  WRFSBASE, WRGSBASE and SWAPGS instructions.)

Thus, document the segment registers with the same "Selector allowing flat
access to all addresses" language on the "Main.asm" level too, since that
is valid for both 32-bit and 64-bit modes.

(Technically, "Main.asm" does not return, but RBP/EBP, passed similarly to
the SEC entry point, is already documented as an output parameter.)

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Suggested-by: Jordan Justen <jordan.l.justen@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
4 years agoArmVirtPkg: switch to new PL011UartLib implementation
Ard Biesheuvel [Thu, 16 Nov 2017 17:08:08 +0000 (17:08 +0000)]
ArmVirtPkg: switch to new PL011UartLib implementation

Switch to the new, cleaned up PL011UartLib implementation so we will
be able to remove the old one.

Contributed-under: TianoCore Contribution Agreement 1.1
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 agoArmPlatformPkg: reorganize PL011 code
Ard Biesheuvel [Thu, 16 Nov 2017 17:08:08 +0000 (17:08 +0000)]
ArmPlatformPkg: reorganize PL011 code

The PL011 code in ArmPlatformPkg is organized in a weird way: there is
a single PL011Uart.h header file under Include/Drivers containing both
register definitions and function entry points. The PL011Uart library
itself is in Drivers/ but it is actually a library.

So let's clean this up: add a new PL011UartLib library class and associated
header file containing only the library prototypes, and move the library
itself under Library/ using a new GUID, with the register definitions moved
into a local header file.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoOvmfPkg/XenHypercallLib: enable virt extensions for ARM
Ard Biesheuvel [Thu, 16 Nov 2017 16:51:49 +0000 (16:51 +0000)]
OvmfPkg/XenHypercallLib: enable virt extensions for ARM

XenHypercallLib uses the 'hvc' instruction, which is not implemented
on all ARMv7 CPUs, and so we need to explicitly specify a CPU that
has the virtualization extensions.

This override used to be set at the platform level, but this was removed
in commit 0d36a219c7bdbb27d775b50837823b2a9928147c
('ArmPlatformPkg/PL031RealTimeClockLib: drop ArmPlatformSysConfigLib
reference), under the assumption that all users of the 'hvc' instruction
had already been fixed.

So fix this for GNU binutils by adding the 'virt' arch extension
directive, and for RVCT by setting the --cpu command line option to a
CPU that is virt capable.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Laszlo Ersek <lersek@redhat.com>
4 years agoMdeModulePkg/SdBlockIoPei: Support IoMmu
Hao Wu [Mon, 30 Oct 2017 05:49:31 +0000 (13:49 +0800)]
MdeModulePkg/SdBlockIoPei: Support IoMmu

Update the SdBlockIoPei driver to consume IOMMU_PPI to allocate DMA
buffer.

If no IOMMU_PPI exists, this driver still calls PEI service
to allocate DMA buffer, with assumption that DRAM==DMA.

This is a compatible change.

Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg/EmmcBlockIoPei: Support IoMmu
Hao Wu [Mon, 30 Oct 2017 03:15:08 +0000 (11:15 +0800)]
MdeModulePkg/EmmcBlockIoPei: Support IoMmu

Update the EmmcBlockIoPei driver to consume IOMMU_PPI to allocate DMA
buffer.

If no IOMMU_PPI exists, this driver still calls PEI service
to allocate DMA buffer, with assumption that DRAM==DMA.

This is a compatible change.

Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg/UfsBlockIoPei: Support IoMmu
Hao Wu [Wed, 18 Oct 2017 05:37:56 +0000 (13:37 +0800)]
MdeModulePkg/UfsBlockIoPei: Support IoMmu

V2 changes:
Resource cleanup logic update in UfsEndOfPei().

V1 history:
Update the UfsBlockIoPei driver to consume IOMMU_PPI to allocate DMA
buffer.

If no IOMMU_PPI exists, this driver still calls PEI service
to allocate DMA buffer, with assumption that DRAM==DMA.

This is a compatible change.

Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg XhciPei: Minor refinement about IoMmu
Star Zeng [Thu, 16 Nov 2017 10:44:07 +0000 (18:44 +0800)]
MdeModulePkg XhciPei: Minor refinement about IoMmu

1. Call IoMmuInit() after locating gPeiUsbControllerPpiGuid.
2. Call XhcPeiFreeSched() to do cleanup in XhcEndOfPei.

Cc: Jiewen Yao <jiewen.yao@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>
4 years agoMdeModulePkg EhciPei: Support IoMmu
Star Zeng [Wed, 15 Nov 2017 05:35:11 +0000 (13:35 +0800)]
MdeModulePkg EhciPei: Support IoMmu

V2: Halt HC at EndOfPei.

Update the EhciPei driver to consume IOMMU_PPI to allocate DMA buffer.

If no IOMMU_PPI exists, this driver still calls PEI service to allocate
DMA buffer, with assumption that DRAM==DMA.

Cc: Jiewen Yao <jiewen.yao@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>
4 years agoMdeModulePkg/PiSmmCore: Implement heap guard feature for SMM mode
Jian J Wang [Tue, 14 Nov 2017 02:56:37 +0000 (10:56 +0800)]
MdeModulePkg/PiSmmCore: Implement heap guard feature for SMM mode

This feature makes use of paging mechanism to add a hidden (not present)
page just before and after the allocated memory block. If the code tries
to access memory outside of the allocated part, page fault exception will
be triggered.

This feature is controlled by three PCDs:

    gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPropertyMask
    gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPoolType
    gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPageType

BIT2 and BIT3 of PcdHeapGuardPropertyMask can be used to enable or disable
memory guard for SMM page and pool respectively. PcdHeapGuardPoolType and/or
PcdHeapGuardPageType are used to enable or disable guard for specific type
of memory. For example, we can turn on guard only for EfiRuntimeServicesCode
and EfiRuntimeServicesData by setting the PCD with value 0x60.

Pool memory is not ususally integer multiple of one page, and is more likely
less than a page. There's no way to monitor the overflow at both top and
bottom of pool memory. BIT7 of PcdHeapGuardPropertyMask is used to control
how to position the head of pool memory so that it's easier to catch memory
overflow in memory growing direction or in decreasing direction.

Cc: Eric Dong <eric.dong@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Suggested-by: Ayellet Wolman <ayellet.wolman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: Add SmmMemoryAttribute protocol
Jian J Wang [Tue, 14 Nov 2017 02:56:14 +0000 (10:56 +0800)]
UefiCpuPkg/PiSmmCpuDxeSmm: Add SmmMemoryAttribute protocol

Heap guard makes use of paging mechanism to implement its functionality. But
there's no protocol or library available to change page attribute in SMM mode.
A new protocol gEdkiiSmmMemoryAttributeProtocolGuid is introduced to make it
happen. This protocol provide three interfaces

struct _EDKII_SMM_MEMORY_ATTRIBUTE_PROTOCOL {
  EDKII_SMM_GET_MEMORY_ATTRIBUTES       GetMemoryAttributes;
  EDKII_SMM_SET_MEMORY_ATTRIBUTES       SetMemoryAttributes;
  EDKII_SMM_CLEAR_MEMORY_ATTRIBUTES     ClearMemoryAttributes;
};

Since heap guard feature need to update page attributes. The page table
should not set to be read-only if heap guard feature is enabled for SMM
mode. Otherwise this feature cannot work.

Cc: Eric Dong <eric.dong@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Suggested-by: Ayellet Wolman <ayellet.wolman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
4 years agoMdeModulePkg/DxeCore: Implement heap guard feature for UEFI
Jian J Wang [Tue, 14 Nov 2017 02:55:26 +0000 (10:55 +0800)]
MdeModulePkg/DxeCore: Implement heap guard feature for UEFI

This feature makes use of paging mechanism to add a hidden (not present)
page just before and after the allocated memory block. If the code tries
to access memory outside of the allocated part, page fault exception will
be triggered.

This feature is controlled by three PCDs:

    gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPropertyMask
    gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPoolType
    gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPageType

BIT0 and BIT1 of PcdHeapGuardPropertyMask can be used to enable or disable
memory guard for page and pool respectively. PcdHeapGuardPoolType and/or
PcdHeapGuardPageType are used to enable or disable guard for specific type
of memory. For example, we can turn on guard only for EfiBootServicesData
and EfiRuntimeServicesData by setting the PCD with value 0x50.

Pool memory is not ususally integer multiple of one page, and is more likely
less than a page. There's no way to monitor the overflow at both top and
bottom of pool memory. BIT7 of PcdHeapGuardPropertyMask is used to control
how to position the head of pool memory so that it's easier to catch memory
overflow in memory growing direction or in decreasing direction.

Note1: Turning on heap guard, especially pool guard, will introduce too many
memory fragments. Windows 10 has a limitation in its boot loader, which
accepts at most 512 memory descriptors passed from BIOS. This will prevent
Windows 10 from booting if heap guard is enabled. The latest Linux
distribution with grub boot loader has no such issue. Normally it's not
recommended to enable this feature in production build of BIOS.

Note2: Don't enable this feature for NT32 emulation platform which doesn't
support paging.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Suggested-by: Ayellet Wolman <ayellet.wolman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
4 years agoMdeModulePkg/DxeIpl: Enable paging for heap guard
Jian J Wang [Tue, 14 Nov 2017 02:53:28 +0000 (10:53 +0800)]
MdeModulePkg/DxeIpl: Enable paging for heap guard

Heap guard feature needs paging to work properly. 64-bit BIOS uses
PcdDxeIplBuildPageTables to control the page table setup. 32-bit BIOS
has to check heap guard feature to decide enabling paging or not.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Suggested-by: Ayellet Wolman <ayellet.wolman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
4 years agoUefiCpuPkg/CpuDxe: Reduce debug message
Jian J Wang [Tue, 14 Nov 2017 02:52:31 +0000 (10:52 +0800)]
UefiCpuPkg/CpuDxe: Reduce debug message

Heap guard feature will frequently update page attributes. The debug message
in CpuDxe driver will slow down the boot performance noticeably. Changing the
debug level to DEBUG_VERBOSE to reduce the message output for normal debug
configuration.

Cc: Eric Dong <eric.dong@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Suggested-by: Ayellet Wolman <ayellet.wolman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
4 years agoMdeModulePkg/SmmMemoryAttribute.h: Add new protocol definitions
Jian J Wang [Tue, 14 Nov 2017 02:51:50 +0000 (10:51 +0800)]
MdeModulePkg/SmmMemoryAttribute.h: Add new protocol definitions

The new protocol gEdkiiSmmMemoryAttributeProtocolGuid is intended for
PiSmmCore to be able to change memory page attributes for the sake of
heap guard feature.

This protocol provides three interfaces to get/set/clear page attribute.

struct _EDKII_SMM_MEMORY_ATTRIBUTE_PROTOCOL {
  EDKII_SMM_GET_MEMORY_ATTRIBUTES       GetMemoryAttributes;
  EDKII_SMM_SET_MEMORY_ATTRIBUTES       SetMemoryAttributes;
  EDKII_SMM_CLEAR_MEMORY_ATTRIBUTES     ClearMemoryAttributes;
};

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>
Suggested-by: Ayellet Wolman <ayellet.wolman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
4 years agoMdeModulePkg/MdeModulePkg.dec,.uni: Add Protocol, PCDs and string tokens
Jian J Wang [Fri, 15 Sep 2017 13:48:06 +0000 (21:48 +0800)]
MdeModulePkg/MdeModulePkg.dec,.uni: Add Protocol, PCDs and string tokens

Add definitions and strings for following new PCDs:

  gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPageType
  gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPoolType
  gEfiMdeModulePkgTokenSpaceGuid.PcdHeapGuardPropertyMask

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Suggested-by: Ayellet Wolman <ayellet.wolman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
4 years agoEmbeddedPkg: add mx66u1g45g nor flash info
Pipat Methavanitpong [Wed, 15 Nov 2017 11:19:32 +0000 (11:19 +0000)]
EmbeddedPkg: add mx66u1g45g nor flash info

Add Macronix MX66U1G45G definition to NorFlashInfoLib

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Pipat Methavanitpong <methavanitpong.pipat@socionext.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Marcin Wojtas <mw@semihalf.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmVirtPkg: remove ArmPlatformSysConfigLib dependency
Ard Biesheuvel [Wed, 15 Nov 2017 13:59:08 +0000 (13:59 +0000)]
ArmVirtPkg: remove ArmPlatformSysConfigLib dependency

Now that the PL031 RTC driver library no longer depends on the ARM
platform specific ArmPlatformSysConfigLib, we no longer need to
implement ArmPlatform.h or have a resolution for that library.
This allows us to get rid of a rather dodgy practice of including
platform headers using compiler flags, which is a bad idea at various
levels.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
4 years agoArmPlatformPkg/PL031RealTimeClockLib: drop ArmPlatformSysConfigLib reference
Ard Biesheuvel [Wed, 15 Nov 2017 13:04:46 +0000 (13:04 +0000)]
ArmPlatformPkg/PL031RealTimeClockLib: drop ArmPlatformSysConfigLib reference

The PL031 driver implements a VExpress/Juno specific hack to set the
battery backed clock in addition to the PL031. However, none of the
remaining VExpress based hardware we support in EDK2 actuall implements
this feature so we can just remove it, and get rid of the cumbersome
dependency on ArmPlatform.h.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPkg: move RVCT PLATFORM_FLAGS override into ArmHvcLib/ArmSmcLib
Ard Biesheuvel [Thu, 16 Nov 2017 14:26:27 +0000 (14:26 +0000)]
ArmPkg: move RVCT PLATFORM_FLAGS override into ArmHvcLib/ArmSmcLib

Currently, each ARM platform built with RVCT that uses ArmHvcLib
or ArmSmcLib needs to specify a CPU target that implements both the
security and virtualization extensions, so that the assembler does
not choke on the 'hvc' and 'smc' instructions in ArmHvcLib/ArmSvcLib.
Let's move these overrides into the module .INFs so we can lift this
requirement at the platform side.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmVirtPkg/ArmVirtQemu: use non-accelerated CopyMem for VariableRuntimeDxe
Ard Biesheuvel [Tue, 14 Nov 2017 10:13:03 +0000 (10:13 +0000)]
ArmVirtPkg/ArmVirtQemu: use non-accelerated CopyMem for VariableRuntimeDxe

The VariableRuntimeDxe driver may use CopyMem () on NOR flash regions,
assuming such regions always have full memory semantics. Given that
those regions cannot be mapped as ordinary memory on ARM (due to the
fact that the NOR flash requires device semantics while in write mode)
this prevents us from using BaseMemoryLibOptDxe in VariableRuntimeDxe,
since it may use unaligned accesses and/or DC ZVA instructions, both
of which are incompatible with mappings using device semantics.

Note that there is no way we can work around this by changing the
mapping type between 'memory' and 'device' when switching from read to
write mode and back, because the runtime mapping is created by the OS,
and cannot be changed at will.

So let's just switch to the unaccelerated version of BaseMemoryLib which
does not have the same problem.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Shannon Zhao <zhaoshenglong@huawei.com>
4 years agoMdeModulePkg/SNP: remove redundant DEBUG print in SNP Transmit.c
Fu Siyuan [Wed, 15 Nov 2017 03:11:20 +0000 (11:11 +0800)]
MdeModulePkg/SNP: remove redundant DEBUG print in SNP Transmit.c

This patch is to remove some redundant DEBUG output in SNP transmit function.
In case of return EFI_NOT_READY in PxeTransmit, the SNP driver is indicate
the caller that the transmit queue is full, it's a very common situation druing
transmit, not a critical error. So the patch move the DEBUG lever to EFI_D_NET.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
4 years agoNetworkPkg: Fix incorrect SizeofHeaders returned from HttpTcpReceiveHeader().
Fu Siyuan [Wed, 15 Nov 2017 02:42:17 +0000 (10:42 +0800)]
NetworkPkg: Fix incorrect SizeofHeaders returned from HttpTcpReceiveHeader().

This patch is to fix a bug that the HttpTcpReceiveHeader() may return incorrect
SizeofHeaders, which will include some already received message-body.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
4 years agoNetworkPkg: Print error message to screen if error occurs during HTTP boot.
Fu Siyuan [Mon, 13 Nov 2017 07:20:37 +0000 (15:20 +0800)]
NetworkPkg: Print error message to screen if error occurs during HTTP boot.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
4 years agoShellPkg: Add error message if failed to place receive token in ping command.
Fu Siyuan [Wed, 15 Nov 2017 04:08:42 +0000 (12:08 +0800)]
ShellPkg: Add error message if failed to place receive token in ping command.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
Reviewed-by: Ye Ting <ting.ye@intel.com>
4 years agoCorebootModulePkg/CbSupportDxe: Remove duplicated IO Space addition
Benjamin You [Tue, 14 Nov 2017 08:41:13 +0000 (16:41 +0800)]
CorebootModulePkg/CbSupportDxe: Remove duplicated IO Space addition

Since UefiCpuPkg's CpuDxe Driver already adds Local Apic's MMIO space to
GCD, CorebootModulePkg's CbSupportDxe should not do this again. Doing this
again causes error return status from GCD service, and ASSERT (FALSE) with
debug build, so the duplicated addition is removed.

Cc: Maurice Ma <maurice.ma@intel.com>
Cc: Prince Agyeman <prince.agyeman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Benjamin You <benjamin.you@intel.com>
Reviewed-by: Maurice Ma <maurice.ma@intel.com>
4 years agoShellPkg/Shell: Check the OpenVolume result in OpenRootByHandle()
Hao Wu [Tue, 14 Nov 2017 05:29:03 +0000 (13:29 +0800)]
ShellPkg/Shell: Check the OpenVolume result in OpenRootByHandle()

REF: https://bugzilla.tianocore.org/show_bug.cgi?id=779

For the API EfiShellOpenRootByHandle():

The return status of the call to SimpleFileSystem->OpenVolume should be
checked.

It is possible that there is a media change in the device (like CD/DVD
ROM). In such case, the volume root opened and/or the device path opened
previously (also within EfiShellOpenRootByHandle) may be invalid.

This commit adds a check for the result of OpenVolume before subsequently
calling functions like EfiShellGetMapFromDevicePath() &
ConvertEfiFileProtocolToShellHandle().

Cc: Jaben Carsey <jaben.carsey@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>
4 years agoMdeModulePkg/PartitionDxe: Fix UDF fs access on certain CD/DVD medias
Paulo Alcantara [Fri, 13 Oct 2017 13:24:49 +0000 (21:24 +0800)]
MdeModulePkg/PartitionDxe: Fix UDF fs access on certain CD/DVD medias

REF: https://bugzilla.tianocore.org/show_bug.cgi?id=725

Historically many drives or medias do not correctly return their value
for block N - which is also referred as last addressable/recorded block.
When they do so, there is still a problem when relying on last recorded
block number returned by SCSI commands READ CAPACITY and READ TRACK
INFORMATION - that is, between block 0 and block N there may be
unwritten blocks which are located outside any track.

That said, the Partition driver was unable to find AVDP at block N on
certain medias that do not either return or report their last recorded
block number correctly.

Apparently there is no official or correct way to find the correct block
number, however tools like the Philips UDF Conformance Tool (udf_test)
apply a correction by searching for an AVDP or VAT in blocks N through
N-456 -- this can be observed by looking at the log reported by udf_test
on those CD/DVD medias. So, if the AVDP or VAT is found, then it sets
the last recorded block number to where AVDP or VAT was located.

With the below setence in UDF 2.60, 6.13.2.2 Background Physical
Formatting:

"... the second AVDP must be recorded after the Background physical
Formatting has been finished..."

Implies that the last recorded block is the one where second AVDP was
recorded.

This patch implements a similar way to correct the last recorded block
number by searching for last AVDP in blocks N-1 through N-512 on those
certain medias, as well as ensure a minimum number of 2 AVDPs found as
specified by ECMA 167 and UDF 2.60 specifications.

Cc: Hao Wu <hao.a.wu@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Reported-by: Hao Wu <hao.a.wu@intel.com>
Signed-off-by: Paulo Alcantara <pcacjr@zytor.com>
Reviewed-by: Hao Wu <hao.a.wu@intel.com>
4 years agoQuarkPlatformPkg/PlatformBootManagerLib: Update boot mode handling
Michael D Kinney [Tue, 7 Nov 2017 22:36:28 +0000 (14:36 -0800)]
QuarkPlatformPkg/PlatformBootManagerLib: Update boot mode handling

https://bugzilla.tianocore.org/show_bug.cgi?id=768

Update and simplify boot mode handling logic.  This includes
performing minimum connects for the following boot modes:

  BOOT_ASSUMING_NO_CONFIGURATION_CHANGES
  BOOT_WITH_MINIMAL_CONFIGURATION
  BOOT_ON_S4_RESUME

Cc: Kelly Steele <kelly.steele@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: Kelly Steele <kelly.steele@intel.com>
4 years agoQuarkPlatformPkg/Readme.md: Fix markdown format issue
Michael D Kinney [Wed, 8 Nov 2017 23:53:03 +0000 (15:53 -0800)]
QuarkPlatformPkg/Readme.md: Fix markdown format issue

https://bugzilla.tianocore.org/show_bug.cgi?id=769

Remove extra carriage return that causes formatting
issues.

Cc: Kelly Steele <kelly.steele@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: Kelly Steele <kelly.steele@intel.com>
4 years agoMdeModulePkg/UsbMassStorageDxe: Fix USB Mass Storage detection
Michael D Kinney [Mon, 30 Oct 2017 20:57:32 +0000 (13:57 -0700)]
MdeModulePkg/UsbMassStorageDxe: Fix USB Mass Storage detection

https://bugzilla.tianocore.org/show_bug.cgi?id=766

Update logic to not return an error from UsbBootRequestSense()
if a Request Sense command responds with no sense information.
It is legal for a USB mass storage device to respond to a
Request Sense command with a SenseKey of
USB_BOOT_SENSE_NO_SENSE and an Additional Sense
Code of USB_BOOT_ASC_NO_ADDITIONAL_SENSE_INFORMATION.

This is described in Section 3.3 of the Universal Serial
Bus Mass Storage Specification For Bootability:

http://www.usb.org/developers/docs/devclass_docs/usb_msc_boot_1.0.pdf

The previous logic returned an error of EFI_NO_RESPONSE
and this caused USB mass storage devices such as a USB
floppy drive to not be detected.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoShellPkg: Fix MSFT C4255 warning
Song, BinX [Fri, 10 Nov 2017 04:11:35 +0000 (12:11 +0800)]
ShellPkg: Fix MSFT C4255 warning

V2:
Fix MSFT C4255 warning
V1:
Enable MSFT C4255 warning.

From MSDN:
Compiler Warning (level 4) C4255
function' : no function prototype given: converting '()' to '(void)'
The compiler did not find an explicit list of arguments to a function.
This warning is for the C compiler only.

Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Bell Song <binx.song@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Jaben Carsey <jaben.carsey@intel.com>
4 years agoNetworkPkg: Fix MSFT C4255 warning
Song, BinX [Fri, 10 Nov 2017 04:11:23 +0000 (12:11 +0800)]
NetworkPkg: Fix MSFT C4255 warning

V2:
Fix MSFT C4255 warning
V1:
Enable MSFT C4255 warning.

From MSDN:
Compiler Warning (level 4) C4255
function' : no function prototype given: converting '()' to '(void)'
The compiler did not find an explicit list of arguments to a function.
This warning is for the C compiler only.

Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Bell Song <binx.song@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdePkg: Fix MSFT C4255 warning
Song, BinX [Fri, 10 Nov 2017 04:11:11 +0000 (12:11 +0800)]
MdePkg: Fix MSFT C4255 warning

V2:
Fix MSFT C4255 warning
V1:
Enable MSFT C4255 warning.

From MSDN:
Compiler Warning (level 4) C4255
function' : no function prototype given: converting '()' to '(void)'
The compiler did not find an explicit list of arguments to a function.
This warning is for the C compiler only.

Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Bell Song <binx.song@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg: Fix MSFT C4255 warning
Song, BinX [Fri, 10 Nov 2017 04:11:02 +0000 (12:11 +0800)]
MdeModulePkg: Fix MSFT C4255 warning

V2:
Fix MSFT C4255 warning
V1:
Enable MSFT C4255 warning.

From MSDN:
Compiler Warning (level 4) C4255
function' : no function prototype given: converting '()' to '(void)'
The compiler did not find an explicit list of arguments to a function.
This warning is for the C compiler only.

Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Bell Song <binx.song@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoIntelFrameworkModulePkg: Fix MSFT C4255 warning
Song, BinX [Fri, 10 Nov 2017 04:10:48 +0000 (12:10 +0800)]
IntelFrameworkModulePkg: Fix MSFT C4255 warning

V2:
Fix MSFT C4255 warning
V1:
Enable MSFT C4255 warning.

From MSDN:
Compiler Warning (level 4) C4255
function' : no function prototype given: converting '()' to '(void)'
The compiler did not find an explicit list of arguments to a function.
This warning is for the C compiler only.

Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Bell Song <binx.song@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoBaseTools: Fix the bug to re-build uni file for Library
Yonghong Zhu [Mon, 13 Nov 2017 02:50:49 +0000 (10:50 +0800)]
BaseTools: Fix the bug to re-build uni file for Library

The root cause is Module's self.CanSkip() is before LibraryAutoGen,
then when a uni file of library is changed, Module's self.CanSkip() is
still true which cause the library is not regenerated.
This patch change Module's self.CanSkip() after LibraryAutoGen.

Fixes: https://bugzilla.tianocore.org/show_bug.cgi?id=759
Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yonghong Zhu <yonghong.zhu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg/Core/Dxe: Remove extra connects for UEFI Applications
Michael D Kinney [Tue, 7 Nov 2017 00:57:36 +0000 (16:57 -0800)]
MdeModulePkg/Core/Dxe: Remove extra connects for UEFI Applications

https://bugzilla.tianocore.org/show_bug.cgi?id=765

The UEFI Specification Boot Services chapter, StartImage() service,
EFF 1.10 Extension requires extra calls to ConnectController()
if a UEFI Driver produces handles. The DXE Core is performing these
extra calls to ConnectController() without evaluating the ImageType.

A filter is added to not make extra calls to ConnectController()
if the ImageType is EFI_IMAGE_SUBSYSTEM_EFI_APPLICATION.

Without this filter, extra calls to ConnectController() may be
performed by UEFI Applications or a UEFI Shell Applications that
also call ConnectController().

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg/TerminalDxe: Fix PCANSI mapping for TRIANGLE and ARROW
Michael D Kinney [Tue, 7 Nov 2017 01:11:17 +0000 (17:11 -0800)]
MdeModulePkg/TerminalDxe: Fix PCANSI mapping for TRIANGLE and ARROW

https://bugzilla.tianocore.org/show_bug.cgi?id=761

When a TerminalType is set to PCANSI, characters in the range 0x00
to 0x1F are control characters.  The mapping table for PCANSI maps
TRIANGLE glyphs, ARROW_UP glyph, and ARROW_DOWN glyph into this
control character range and that causes no characters to be
displayed by PCANSI compatible terminal emulators.

The mappings are updated so these glyphs are mapped to ANSI
characters in the range 0x20 to 0x7E.

GEOMETRICSHAPE_UP_TRIANGLE     '^'
GEOMETRICSHAPE_RIGHT_TRIANGLE  '>'
GEOMETRICSHAPE_DOWN_TRIANGLE   'v'
GEOMETRICSHAPE_LEFT_TRIANGLE   '<'
ARROW_UP                       '^'
ARROW_DOWN                     'v'

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Ruiyu Ni <Ruiyu.ni@intel.com>
4 years agoStdLib/BsdSocketLib: Remove unused variables
Thiebaud Weksteen [Thu, 26 Oct 2017 09:58:15 +0000 (02:58 -0700)]
StdLib/BsdSocketLib: Remove unused variables

TianoCore BZ#677

Remove unused variables and turn nsdispatch macro into a function. This
is not what is recommended in the bug ticket but avoid heavier
modification of the code (and deviation from upsteam).

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Thiebaud Weksteen <tweek@google.com>
Reviewed-by: Jaben Carsey <jaben.carsey@intel.com>
4 years agoArmPlatformPkg/PL031RealTimeClockLib: ignore DST setting when timezone is set
Ard Biesheuvel [Mon, 6 Nov 2017 18:08:15 +0000 (18:08 +0000)]
ArmPlatformPkg/PL031RealTimeClockLib: ignore DST setting when timezone is set

According to the UEFI spec, the timezone setting which the platform needs
to record in addition to the actual date and time already reflects the
current DST setting. In other words, moving the clock from standard time
to daylight saving time also involves adding or subtracting 60 minutes
from the timezone setting, as well as flicking the EFI_TIME_IN_DAYLIGHT
bit in the DST setting.

This means we need to disregard the DST setting if the timezone is
specified, and only add or subtract the additional hour if we are on
local time.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPlatformPkg/PL031RealTimeClockLib: remove validation and DST handling
Ard Biesheuvel [Sun, 5 Nov 2017 16:31:23 +0000 (16:31 +0000)]
ArmPlatformPkg/PL031RealTimeClockLib: remove validation and DST handling

This library, which is intended to encapsulate the hardware specifics
of the ARM PL031 RTC, also implements its own input validation routines
and record the timezone and DST settings in its own set of EFI variables.

This functionality has recently been added to the core driver, so let's
remove it here.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoEmbeddedPkg/RealTimeClockRuntimeDxe: move common functionality into core
Ard Biesheuvel [Sun, 5 Nov 2017 16:06:17 +0000 (16:06 +0000)]
EmbeddedPkg/RealTimeClockRuntimeDxe: move common functionality into core

RealTimeClockRuntimeDxe defers the hardware/platform specific handling
of reading/setting the hardware clock to RealTimeClockLib, but for
unknown reasons, it also defers common functionality such as input
validation and recording the timezone and DST settings (which are
informational only and not managed by hardware)

This has led to a lot of duplication in implementations of RealTimeClockLib
as well as TimeBaseLib, to the point where each library implementation
has its own set of UEFI variables to record the timezone and DST settings.
This makes little sense, and so let's update RealTimeClockRuntimeDxe now
to allow future implementations to rely on the core driver to take care of
these things.

Note that reading the timezone and DST settings occurs before calling into
the library, so we can phase out this behavior gradually from library
implementations in EDK2, edk2-platforms or out of tree.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoMdeModulePkg: Update IP4 stack to support point-to-point link with 31-bit mask.
Fu Siyuan [Wed, 18 Oct 2017 02:13:39 +0000 (10:13 +0800)]
MdeModulePkg: Update IP4 stack to support point-to-point link with 31-bit mask.

V2 update: Directly use NetIp4IsUnicast() to check station address in IP driver.

This patch is to follow RFC3021 which allows to use 31-bit mask
in point-to-point link.
If a 31-bit subnet mask is assigned to a point-to-point link, it
leaves the <Host-number> with only 1 bit.  Consequently, only two
possible addresses may result:
  {<Network-number>, 0} and {<Network-number>, -1}
These addresses have historically been associated with network and
broadcast addresses (see Section 2.2).  In a point-to-point link with
a 31-bit subnet mask, the two addresses above MUST be interpreted as
host addresses.

Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Wu Jiaxin <jiaxin.wu@intel.com>
4 years agoMdePkg: Add definitions for the SPI protocols introduced in PI 1.6.
Marvin.Haeuser@outlook.com [Tue, 31 Oct 2017 00:12:22 +0000 (08:12 +0800)]
MdePkg: Add definitions for the SPI protocols introduced in PI 1.6.

This commit adds header files for the SPI protocols introduced in the
UEFI PI 1.6 specification, as well as their GUIDs to MdePkg.dec.

EFI_SPI_TRANSACTION_TYPE assumes an enum with its members ordered the
way they are listed in the specification, as there are no values given
explicitely.
EFI_LEGACY_SPI_CONTROLLER_GUID assumes the character 'l' used in the
specification was meant to be '1'.

V2:
- Fixed LegacySpiFlash.h's header guard.
- Fixed LegacySpiSmmController's GUID name.
- Removed EFI_SPI_NOR_FLASH_PROTOCOL_LF_READ_DATA as it's unused.
- Added the missing SpiSmmNorFlash header.
- Fixed all file endings to be CRLF.
- Removed trailing whitespaces.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Marvin Haeuser <Marvin.Haeuser@outlook.com>
Reviewed-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoArmPlatformPkg/PrePeiCore: seed temporary stack before entering PEI core
Ard Biesheuvel [Fri, 20 Oct 2017 10:44:04 +0000 (11:44 +0100)]
ArmPlatformPkg/PrePeiCore: seed temporary stack before entering PEI core

DEBUG builds of PEI code will print a diagnostic message regarding
the utilization of temporary RAM before switching to permanent RAM.
For example,

  Total temporary memory:    16352 bytes.
    temporary memory stack ever used:       4820 bytes.
    temporary memory heap used for HobList: 4720 bytes.

Tracking stack utilization like this requires the stack to be seeded
with a known magic value, and this needs to occur before entering C
code, given that it uses the stack. Currently, only Nt32Pkg appears
to implement this feature, but it is useful nonetheless, so let's
wire it up for PrePeiCore.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Liming Gao <liming.gao@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoMdePkg/NvdimmLabel.h: Add NVDIMM_LABEL protocol definition
Ruiyu Ni [Wed, 1 Nov 2017 09:34:52 +0000 (17:34 +0800)]
MdePkg/NvdimmLabel.h: Add NVDIMM_LABEL protocol definition

NVDIMM_LABEL protocol is defined in UEFI 2.7 spec, to provide
services that allow management of labels contained in a Label
Storage Area in NVDIMM.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdePkg/Btt.h: Add Block Translation Table definitions
Ruiyu Ni [Wed, 1 Nov 2017 05:59:38 +0000 (13:59 +0800)]
MdePkg/Btt.h: Add Block Translation Table definitions

BTT definitions are defined in UEFI spec 2.7, to defines
a layout and set of rules for doing block I/O that provide
powerfail write atomicity of a single block in NVDIMM.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg SerialDxe: Handle Timeout change more robustly
Star Zeng [Fri, 27 Oct 2017 03:14:11 +0000 (11:14 +0800)]
MdeModulePkg SerialDxe: Handle Timeout change more robustly

https://lists.01.org/pipermail/edk2-devel/2017-October/016479.html
reported "Xen Console input very slow in recent UEFI" that appears
after 4cf3f37c87ba1f9d58072444bd735e40e4779e70 "MdeModulePkg
SerialDxe: Process timeout consistently in SerialRead".

Julien did more debugging and find out the following is happening in
TerminalConInTimerHandler (MdeModulePkg/Universal/Console/TerminalDxe)
when a character is received:
1) GetControl will return EFI_SERIAL_INPUT_BUFFER_EMPTY unset
  => Entering in the loop to fetch character from the serial
2) GetOneKeyFromSerial()
  => Return directly with the character read
3) Looping as the fifo is not full and no error
4) GetOneKeyFromSerial() -> SerialRead()
  => No more character so SerialPortPoll() will return FALSE and loop
     until timeout
  => Return EFI_TIMEOUT
5) Exiting the loop from TerminalConInTimerHandler
6) Characters are printed

After some investigation, I found it is related to the Timeout value.

The Timeout is 1000000 (1s) by default to follow UEFI spec.
And the Terminal driver will recalculate and set the Timeout value
based on the properties of UART in TerminalDriverBindingStart()/
TerminalConInTimerHandler().

  SerialInTimeOut = 0;
  if (Mode->BaudRate != 0) {
    //
    // According to BAUD rate to calculate the timeout value.
    //
    SerialInTimeOut = (1 + Mode->DataBits + Mode->StopBits) *
                      2 * 1000000 / (UINTN) Mode->BaudRate;
  }

For example, based on the PCD values of PcdUartDefaultBaudRate,
PcdUartDefaultDataBits and PcdUartDefaultStopBits, SerialInTimeOut =
(1 + 8  + 1) * 2 * 1000000 / (UINTN) 115200 = 173 (us).

When SerialDxe is used,
TerminalDriverBindingStart()/TerminalConInTimerHandler() ->
  SerialIo->SetAttributes() ->
    SerialSetAttributes() ->
      SerialPortSetAttributes()

Some implementations of SerialPortSetAttributes() could handle the
input parameters and return RETURN_SUCCESS, for example
BaseSerialPortLib16550, then Timeout value will be changed to 173 (us),
no "slow down" will be observed.
But some implementations of SerialPortSetAttributes() just return
RETURN_UNSUPPORTED, for example XenConsoleSerialPortLib, then Timeout
value will be not changed and kept 1000000 (1s), "slow down" will be
observed.

SerialPortLib instance can be enhanced to
1. Handle the input parameters and return status accordingly instead of
just returning RETURN_UNSUPPORTED in SerialPortSetAttributes().
2. Just return RETURN_SUCCESS instead of RETURN_UNSUPPORTED in
SerialPortSetAttributes() if the instance does not care the input
parameters at all.

And SerialDxe can also be enhanced like this patch to be more robust
to handle Timeout change.

Cc: Julien Grall <julien.grall@linaro.org>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Compare against the original parameters
Suggested-by: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Julien Grall <julien.grall@linaro.org>
4 years agoIntelFsp2Pkg-Tools: GenCfgOpt.py shouldn't include specific UPD name
Chasel, Chiu [Wed, 8 Nov 2017 12:35:26 +0000 (20:35 +0800)]
IntelFsp2Pkg-Tools: GenCfgOpt.py shouldn't include specific UPD name

PcdSerialIoUartDebugEnable UPD is platform specific and should not
be included in generic GenCfgOpt.py script. Remove this and platform
DSC should control the default value instead.

Cc: Jiewen Yao <Jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Chasel Chiu <chasel.chiu@intel.com>
Reviewed-by: Jiewen Yao <Jiewen.yao@intel.com>
4 years agoIntelFrameworkModulePkg: Fix misuses of AllocateCopyPool
Jian J Wang [Wed, 8 Nov 2017 02:09:54 +0000 (10:09 +0800)]
IntelFrameworkModulePkg: Fix misuses of AllocateCopyPool

AllocateCopyPool(AllocationSize, *Buffer) will copy "AllocationSize" bytes of
memory from old "Buffer" to new allocated one. If "AllocationSize" is bigger
than size of "Buffer", heap memory overflow occurs during copy.

One solution is to allocate pool first then copy the necessary bytes to new
memory. Another is using ReallocatePool instead if old buffer will be freed
on spot.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Bi Dandan <dandan.bi@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Bi Dandan <dandan.bi@intel.com>
4 years agoShellPkg: Fix misuses of AllocateCopyPool
Jian J Wang [Wed, 8 Nov 2017 02:09:19 +0000 (10:09 +0800)]
ShellPkg: Fix misuses of AllocateCopyPool

AllocateCopyPool(AllocationSize, *Buffer) will copy "AllocationSize" bytes of
memory from old "Buffer" to new allocated one. If "AllocationSize" is bigger
than size of "Buffer", heap memory overflow occurs during copy.

One solution is to allocate pool first then copy the necessary bytes to new
memory. Another is using ReallocatePool instead if old buffer will be freed
on spot.

Cc: Jaben Carsey <jaben.carsey@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Bi Dandan <dandan.bi@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>
4 years agoMdeModulePkg: Fix misuses of AllocateCopyPool
Jian J Wang [Wed, 1 Nov 2017 15:18:34 +0000 (23:18 +0800)]
MdeModulePkg: Fix misuses of AllocateCopyPool

AllocateCopyPool(AllocationSize, *Buffer) will copy "AllocationSize" bytes of
memory from old "Buffer" to new allocated one. If "AllocationSize" is bigger
than size of "Buffer", heap memory overflow occurs during copy.

One solution is to allocate pool first then copy the necessary bytes to new
memory. Another is using ReallocatePool instead if old buffer will be freed
on spot.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Bi Dandan <dandan.bi@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Bi Dandan <dandan.bi@intel.com>
4 years agoShellPkg/dh: Fix wrong output when dumping PciRootBridgeIo
Ruiyu Ni [Tue, 7 Nov 2017 07:34:23 +0000 (15:34 +0800)]
ShellPkg/dh: Fix wrong output when dumping PciRootBridgeIo

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Jaben Carsey <jaben.carsey@intel.com>
4 years agoMdeModulePkg/NonDiscoverable: fix memory override bug
Heyi Guo [Mon, 30 Oct 2017 05:47:52 +0000 (13:47 +0800)]
MdeModulePkg/NonDiscoverable: fix memory override bug

For PciIoPciRead interface, memory prior to Buffer would be written
with zeros if Offset was larger than sizeof (Dev->ConfigSpace), which
would cause serious system exception.

So we add a pre-check branch to avoid memory override.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Heyi Guo <heyi.guo@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
4 years agoMdePkg/PciExpress21.h: Fix typo in PCI_REG_PCIE_SLOT_CONTROL
Ruiyu Ni [Thu, 26 Oct 2017 08:06:56 +0000 (16:06 +0800)]
MdePkg/PciExpress21.h: Fix typo in PCI_REG_PCIE_SLOT_CONTROL

PCI_REG_PCIE_SLOT_CONTROL contains a typo. It is defined as:
typedef union {
  struct {
    UINT32 AttentionButtonPressed : 1;
    UINT32 ...
    ...
  } Bits;
  UINT16   Uint16;
} PCI_REG_PCIE_SLOT_CONTROL;

The bit field data type should be UINT16 instead of UINT32,
results sizeof (PCI_REG_PCIE_SLOT_CONTROL) equals to 4 instead of 2.

Because this structure is used in PCI_CAPABILITY_PCIEXP as below:
typedef struct {
  ...
  PCI_REG_PCIE_SLOT_CONTROL       SlotControl;
  PCI_REG_PCIE_SLOT_STATUS        SlotStatus;
} PCI_CAPABILITY_PCIEXP;

It cause the OFFSET_OF (PCI_CAPABILITY_PCIEXP, SlotStatus) equal
to a wrong value.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoPcAtChipsetPkg/IsaAcpiDxe: Restore PCI attributes correctly
Ruiyu Ni [Thu, 2 Nov 2017 01:21:56 +0000 (09:21 +0800)]
PcAtChipsetPkg/IsaAcpiDxe: Restore PCI attributes correctly

REF: https://bugzilla.tianocore.org/show_bug.cgi?id=405

The original code enables some BITs in PCI attributes in Start(),
but wrongly to disable these BITs in Stop().

The correct behavior is to save the original PCI attributes before
enables some BITs in Start(), and restore to original value
in Stop().

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
4 years agoEmbeddedPkg: Implement NorFlashInfoLib
Marcin Wojtas [Fri, 3 Nov 2017 17:55:17 +0000 (18:55 +0100)]
EmbeddedPkg: Implement NorFlashInfoLib

The SPI NOR flash drivers which base on ArmPlatformPkg's
NorFlashDxe usually make use of static declarations of the
flash instances with their type and parameters. As a result
it implies hardcoding the exact way of flash handling, not to
mention the code does not look very nice. Much better solution
would be obtaining the flash ID and hence its description
in runtime.

JEDEC compliant SPI NOR devices allow to obtain their IDs with
READ_ID command (0x9f), which should return the vendor ID byte,
followed by 2 to 4 following device ID bytes. Use this capability
for implementing a NorFlashInfoLib that gives an access to the
NOR flash description data, such as name, page size, sector
(block) size and others, of more than 50 different models.
The new library user should pass an output array from issuing
READ_ID command to the NorFlashGetInfo () routine - if the
match is found, an allocated (optionally for RT) pool with
the flash description copy will be returned.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoSecurityPkg/AuthVariableLib: Use EFI_CERT_DATA to parse certificate
chenc2 [Tue, 7 Nov 2017 01:01:26 +0000 (09:01 +0800)]
SecurityPkg/AuthVariableLib: Use EFI_CERT_DATA to parse certificate

The function Pkcs7GetSigners return certificate stack as binary buffer.
Use EFI_CERT_DATA to parsing certificate stack more clearly, and access
certificate by the field of EFI_CERT_DATA structure.

Cc: Long Qin <qin.long@intel.com>
Cc: Zhang Chao <chao.b.zhang@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: chenc2 <chen.a.chen@intel.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Zhang Chao <chao.b.zhang@intel.com>
4 years agoCryptoPkg/BaseCryptLib: Add C-structure to matching certificate stack
chenc2 [Tue, 7 Nov 2017 00:56:56 +0000 (08:56 +0800)]
CryptoPkg/BaseCryptLib: Add C-structure to matching certificate stack

The parameter CertStack of Pkcs7GetSigners will return all embedded X.509
certificate in one given PKCS7 signature. The format is:
//
// UINT8  CertNumber;
// UINT32 Cert1Length;
// UINT8  Cert1[];
// UINT32 Cert2Length;
// UINT8  Cert2[];
// ...
// UINT32 CertnLength;
// UINT8  Certn[];
//
Add EFI_CERT_STACK and EFI_CERT_DATA structure, these two C-structure are
used for parsing CertStack more clearly.

Cc: Long Qin <qin.long@intel.com>
Cc: Zhang Chao <chao.b.zhang@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: chenc2 <chen.a.chen@intel.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Zhang Chao <chao.b.zhang@intel.com>
4 years agoArmPkg/ArmMmuLib: Add new attribute WRITE_BACK_NONSHAREABLE
Peicong Li [Tue, 7 Nov 2017 12:56:51 +0000 (20:56 +0800)]
ArmPkg/ArmMmuLib: Add new attribute WRITE_BACK_NONSHAREABLE

Flash region needs to be set as cacheable (write back) to increase
performance, if PEI is still XIP on flash or DXE FV is decompressed
from flash FV. However some ARM platforms do not support to set flash
as inner shareable since flash is not normal DDR memory and it will
not respond to cache snoop request, which will causes system hang
after MMU is enabled.

So we need a new ARM memory region attribute WRITE_BACK_NONSHAREABLE
for flash region on these platforms specifically. This attribute will
set the region as write back but not inner shared.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Peicong Li <lipeicong@huawei.com>
Signed-off-by: Heyi Guo <heyi.guo@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoMdeModulePkg/PciBus: Disable BME of all devices when entering RT
Ruiyu Ni [Tue, 31 Oct 2017 07:53:18 +0000 (15:53 +0800)]
MdeModulePkg/PciBus: Disable BME of all devices when entering RT

The patch ensures all DMA transactions are blocked after
ExitBootService.
If a platform enables IOMMU before and needs disable IOMMU after
ExitBootService, the IOMMU should be disabled after PCI bus driver
disables BME.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael Turner <michael.turner@microsoft.com>
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Cc: Jeff Fan <vanjeff_919@hotmail.com>
4 years agoShellPkg/HexEdit: Fix EBC build failure
Ruiyu Ni [Fri, 3 Nov 2017 08:55:43 +0000 (16:55 +0800)]
ShellPkg/HexEdit: Fix EBC build failure

EfiCpuIoWidthUint8 should be used in HMemImageRead and HMemImageSave.
Because CpuIo protocol is now used for memory access.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Jaben Carsey <jaben.carsey@intel.com>
4 years agoMdeModulePkg/VarCheckHii: Enhance VarCheckHiiLib to support bit check
Dandan Bi [Fri, 3 Nov 2017 08:34:25 +0000 (16:34 +0800)]
MdeModulePkg/VarCheckHii: Enhance VarCheckHiiLib to support bit check

V3:
(1) Reset QuestionStoredInBitField to FALSE at end opcode(EFI_IFR_END_OP)
(2) Fix typo and format issues(line alignment for debug print message
and value assignment...)

V2:
(1)Remove the VarOffsetBitLevel/StorageWidthBitLevel to reduce the final
VarCheckBinSize and update the implementation accordingly.
(2)Update the VAR_CHECK_HII_REVISION
(3)Refine the Debug message and function comments,like update oneof",
"checkbox", "numeric" to "OneOf", "CheckBox", "Numeric".

VarCheckHiiLib check the value set to storage based on the possible value
listed in the vfr file. Since we have enhanced vfr to support Question
value stored in bit field, so now enhance VarCheckHiiLib to support bit
field check.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg/VarCheckHiiLib: Replace EFI_D_INFO with DEBUG_INFO
Dandan Bi [Fri, 3 Nov 2017 06:14:02 +0000 (14:14 +0800)]
MdeModulePkg/VarCheckHiiLib: Replace EFI_D_INFO with DEBUG_INFO

In patch 2, we will introduce DEBUG_INFO in VarCheckHiiLib,in order to keep
consistence, replace all EFI_D_INFO with DEBUG_INFO firstly in this pacth.

Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg VarCheckHiiLib: Fix typo "0x02%x" to be "0x%02x"
Star Zeng [Mon, 6 Nov 2017 07:15:55 +0000 (15:15 +0800)]
MdeModulePkg VarCheckHiiLib: Fix typo "0x02%x" to be "0x%02x"

Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Dandan Bi <dandan.bi@intel.com>
4 years agoMdeModulePkg/SdMmcPciHcDxe: call SdMmcFreeTrb() to complete sync operation
Ard Biesheuvel [Sat, 4 Nov 2017 21:32:37 +0000 (21:32 +0000)]
MdeModulePkg/SdMmcPciHcDxe: call SdMmcFreeTrb() to complete sync operation

Currently, we complete a synchronous operation without unmapping the
DMA mappings, and free the pages using FreePages () rather than calling
EFI_PCI_IO_PROTOCOL::FreeBuffer. This is simply incorrect, but it also
breaks non-coherent DMA as well as DMA protection and/or memory encryption
so let's do it correctly and call SdMmcFreeTrb() instead.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoCryptoPkg/BaseCryptLib: Fix mismatched memory allocation/free
Long Qin [Tue, 31 Oct 2017 07:50:30 +0000 (15:50 +0800)]
CryptoPkg/BaseCryptLib: Fix mismatched memory allocation/free

The malloc/free (instead of AllocatePool/FreePool) were used directly
in some wrapper implementations, which was designed to leverage the
light-weight memory management routines at Runtime phase.
The malloc/free and AllocatePool/FreePool usages are required to be
matched, after extra memory size info header was introduced in malloc
wrapper.

This patch corrects two memory allocation cases, which requires the
caller to free the buffer with FreePool() outside the function call.

And some comments were also added to clarify the correct memory
release functions if it's the caller's responsibility to free the
memory buffer.

Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ting Ye <ting.ye@intel.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Jian J Wang <jian.j.wang@intel.com>
4 years agoCryptoPkg/BaseCryptLib: Fix buffer overflow issue in realloc wrapper
Long Qin [Wed, 1 Nov 2017 08:10:04 +0000 (16:10 +0800)]
CryptoPkg/BaseCryptLib: Fix buffer overflow issue in realloc wrapper

There is one long-standing problem in CRT realloc wrapper, which will
cause the obvious buffer overflow issue when re-allocating one bigger
memory block:
    void *realloc (void *ptr, size_t size)
    {
      //
      // BUG: hardcode OldSize == size! We have no any knowledge about
      // memory size of original pointer ptr.
      //
      return ReallocatePool ((UINTN) size, (UINTN) size, ptr);
    }
This patch introduces one extra header to record the memory buffer size
information when allocating memory block from malloc routine, and re-wrap
the realloc() and free() routines to remove this BUG.

Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ting Ye <ting.ye@intel.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Qin Long <qin.long@intel.com>
Reviewed-by: Jian J Wang <jian.j.wang@intel.com>
Validated-by: Jian J Wang <jian.j.wang@intel.com>
4 years agoIntelSiliconPkg IntelVTdDxe: Use ACPI table event to get DMAR table
Star Zeng [Wed, 25 Oct 2017 09:23:03 +0000 (17:23 +0800)]
IntelSiliconPkg IntelVTdDxe: Use ACPI table event to get DMAR table

Use ACPI table event to get DMAR table instead of using ACPI SDT
notification as ACPI SDT is optional and the default value of
PcdInstallAcpiSdtProtocol is FALSE in MdeModulePkg.dec.

Cc: Jiewen Yao <jiewen.yao@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>
4 years agoArmPlatformPkg/PlatformPeim: allow PlatformPeiLib to set the boot mode
Ard Biesheuvel [Wed, 1 Nov 2017 12:57:26 +0000 (12:57 +0000)]
ArmPlatformPkg/PlatformPeim: allow PlatformPeiLib to set the boot mode

The current interdepencies between the PrePeiCore SEC module, the
platform PEIM and ArmPlatformLib is a bit awkward: due to the fact
that ArmPlatformLib is also used by SEC modules, we cannot use PEI
specific facilities in the implementation of ArmPlatformGetBootMode.
However, given that we call that library function /after/ invoking
PlatformPeiLib, there is no way for that library to set the boot mode
other than resorting to tricks like notification callbacks on arbitrary
unrelated events.

ArmPlatformLib should probably be phased out anyway, given its quirky
nature, but for now, let's fix this particular issue by deferring the
call to PlatformPeim() to after the point where we set the boot mode
by calling ArmPlatformGetBootMode ().

While we're at it, clean up the code slightly by using PeiServicesLib
instead of doing double pointer dereferencing.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoEmbeddedPkg: introduce GPIO PPI
Ard Biesheuvel [Wed, 1 Nov 2017 12:54:39 +0000 (12:54 +0000)]
EmbeddedPkg: introduce GPIO PPI

Introduce a PPI counterpart of the existing 'embedded GPIO' protocol,
so we can manipulate GPIOs from PEI modules. This allows things like
setting the boot mode based on a DIP switch setting.

Note that the naming is slightly awkward, as there is nothing 'embedded'
about a GPIO, but given that the DXE protocol already resides here and
has the 'embedded' prefix, it makes sense to retain uniformity.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoNt32Pkg: Update SecMain consumes PCD to get the init value in temp stack
Liming Gao [Thu, 2 Nov 2017 02:59:15 +0000 (10:59 +0800)]
Nt32Pkg: Update SecMain consumes PCD to get the init value in temp stack

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg: Update PeiCore consumes PCD to get the init value in temp stack
Liming Gao [Thu, 2 Nov 2017 02:59:14 +0000 (10:59 +0800)]
MdeModulePkg: Update PeiCore consumes PCD to get the init value in temp stack

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg: Add new PCD PcdInitValueInTempStack
Liming Gao [Thu, 2 Nov 2017 02:59:13 +0000 (10:59 +0800)]
MdeModulePkg: Add new PCD PcdInitValueInTempStack

https://bugzilla.tianocore.org/show_bug.cgi?id=740

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
4 years agoBaseTools: parse map file generated by Xcode on Mac
Yonghong Zhu [Thu, 2 Nov 2017 05:15:34 +0000 (13:15 +0800)]
BaseTools: parse map file generated by Xcode on Mac

Add support to parse map file generated by Xcode on Mac to get
variable offset and Patchable Pcd info in current EFI file.

Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yonghong Zhu <yonghong.zhu@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoBaseTools/VfrCompile: Add check to avoid using NULL pointer
Dandan Bi [Wed, 1 Nov 2017 08:11:47 +0000 (16:11 +0800)]
BaseTools/VfrCompile: Add check to avoid using NULL pointer

Question value are stored in one specified storage, but the Data type
of the storage is not specified or there is no sub fields in the Data
type sometimes, so we need to add check before using related pointers.
Here list some NULL cases:
(1)For an efivastore which doesn't specify a data structure or a
   data type(UINT8,UINT16...)as the storage, just has VarName and
   VarSize instead, we can not get its data type before parsing
   its VarSize.

(2)For efivastore which just specifies the data type(UINT8,UINT16...)
   not a structure as the storage,this data type doesn't have sub-fields.

Cc: Eric Dong <eric.dong@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoAdd one maintainer to CorebootModulePkg and CorebootPayloadPkg
Benjamin You [Tue, 31 Oct 2017 05:41:50 +0000 (13:41 +0800)]
Add one maintainer to CorebootModulePkg and CorebootPayloadPkg

Add one maintainer (Benjamin You) to CorebootModulePkg and CorebootPayloadPkg

Cc: Maurice Ma <maurice.ma@intel.com>
Cc: Prince Agyeman <prince.agyeman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Benjamin You <benjamin.you@intel.com>
Reviewed-by: Maurice Ma <maurice.ma@intel.com>
4 years agoUefiCpuPkg/MpInitLib:Make function comments align with function
Dandan Bi [Wed, 1 Nov 2017 01:36:53 +0000 (09:36 +0800)]
UefiCpuPkg/MpInitLib:Make function comments align with function

Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Dandan Bi <dandan.bi@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
4 years agoShellPkg/hexeditor: Use CpuIo for memory access
Ruiyu Ni [Wed, 1 Nov 2017 10:12:53 +0000 (18:12 +0800)]
ShellPkg/hexeditor: Use CpuIo for memory access

The original code uses PciRootBridgeIo for memory access.
It worked before MdeModulePkg/PciHostBridgeDxe driver was checked in.
But MdeModulePkg/PciHostBridgeDxe adds checks to ensure the MMIO
access request is in the scope of the current RootBridgeIo instance.
It causes "hexeditor -m 0 2" reports error because memory address 0
surely is not in the scope of any RootBridgeIo instance.
In fact only accessing the MMIO space occupied by the RootBridgeIo
can work.

The patch changes hexeditor to use CpuIo for memory access.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Jaben Carsey <jaben.carsey@intel.com>
4 years agoBaseTools/tools_def AARCH64 ARM: disable PIE linking for .aslc sources
Ard Biesheuvel [Wed, 1 Nov 2017 14:58:42 +0000 (14:58 +0000)]
BaseTools/tools_def AARCH64 ARM: disable PIE linking for .aslc sources

Commit 1894a7c64c0a ("BaseTools/tools_def AARCH64 ARM: disable PIE
linking") works around an issue that was caught due to the fact that
PIE linking produces broken .acpi files. However, v2 of that fix
inadvertently only applied the workaround to the normal linker command
line, and not to the ASLD one, so the issue still persists.

So add the missing -no-pie options for ASLD on ARM and AARCH64.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdeModulePkg: Deprecate EFI_VARIABLE_AUTHENTICATED_WRITE_ACCESS
Zhang, Chao B [Tue, 24 Oct 2017 07:04:00 +0000 (15:04 +0800)]
MdeModulePkg: Deprecate EFI_VARIABLE_AUTHENTICATED_WRITE_ACCESS

Mark EFI_VARIABLE_AUTHENTICATED_WRITE_ACCESS as deprecated.
1. Make SetVariable/QueryVariableInfo return EFI_UNSUPPORTED with this
   attribute
2. No change to GetVariable/GetNextVariableName
Also update several function descriptors accordingly

Cc: Long Qin <qin.long@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Chao Zhang <chao.b.zhang@intel.com>
Reviewed-by: Long Qin <qin.long@intel.com>
4 years agoMdePkg: Deprecate EFI_VARIABLE_AUTHENTICATED_WRITE_ACCESS
Zhang, Chao B [Tue, 24 Oct 2017 06:53:52 +0000 (14:53 +0800)]
MdePkg: Deprecate EFI_VARIABLE_AUTHENTICATED_WRITE_ACCESS

Mark EFI_VARIABLE_AUTHENTICATED_WRITE_ACCESS as deprecated. Also update
some function descriptors accordingly.

Cc: Long Qin <qin.long@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Chao Zhang <chao.b.zhang@intel.com>
Reviewed-by: Long Qin <qin.long@intel.com>
4 years agoSecurityPkg: Remove Counter Based AuthVariable support
Zhang, Chao B [Tue, 24 Oct 2017 06:52:10 +0000 (14:52 +0800)]
SecurityPkg: Remove Counter Based AuthVariable support

Remove counter based auth variable support. also modify several function
descriptors to accommodate the change

Cc: Long Qin <qin.long@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Chao Zhang <chao.b.zhang@intel.com>
Reviewed-by: Long Qin <qin.long@intel.com>