mirror_edk2.git
4 years agoMdeModulePkg/FPDT: Add error message for unsupported case
Dandan Bi [Tue, 10 Apr 2018 05:51:08 +0000 (13:51 +0800)]
MdeModulePkg/FPDT: Add error message for unsupported case

We have updated performance infrastructure in previous commits:
between
https://github.com/tianocore/edk2/commit/73fef64f14d1b97ae9bd4705df3becc022391eba
and
https://github.com/tianocore/edk2/commit/115eae650bfd2be2c2bc37360f4a755065e774c4
Update FPDT drivers to collect the performance data reported by
gEdkiiFpdtExtendedFirmwarePerformanceGuid.
The old implementation which collected performance data through
gEfiFirmwarePerformanceGuid is not supported now.
We should add error message to remind user for this unsupported
case in case anyone use it by mistake.

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 agoSignedCapsulePkg SystemFirmwareUpdateDxe: Fix failure caused by d69d922
Star Zeng [Fri, 13 Apr 2018 09:55:14 +0000 (17:55 +0800)]
SignedCapsulePkg SystemFirmwareUpdateDxe: Fix failure caused by d69d922

d69d9227d046211265de1fab5580c50a65944614 caused system firmware update
failure. It is because FindMatchingFmpHandles() is expected to return
handles matched, but the function returns all handles found.

This patch is to fix the issue.
This patch also assigns mSystemFmpPrivate->Handle for "case 1:" path
in case the Handle is needed by other place in future.

Cc: Michael D Kinney <michael.d.kinney@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: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Michael D Kinney <michael.d.kinney@intel.com>
4 years agoCryptoPkg/OpensslLib: Update OpenSSL version to 1.1.0h
Long Qin [Thu, 12 Apr 2018 02:58:45 +0000 (10:58 +0800)]
CryptoPkg/OpensslLib: Update OpenSSL version to 1.1.0h

(https://bugzilla.tianocore.org/show_bug.cgi?id=927)

(V2 Update:
    Removing the wrong "--remote" option from git submodule update
    command in this commit message. Thanks Laszlo's clarification
    to correct this)

Update OpenSSL version to 1.1.0h release (27-Mar-2018) to include the
fix for CVE-2018-0739 issue (Handling of crafted recursive ASN.1
structures can cause a stack overflow and resulting denial of service,
Refer to https://www.openssl.org/news/secadv/20180327.txt for more
information).

Please note "git pull" will not update the submodule repository.
use the following commend to make your existing submodule track this
update:
   $ git submodule update --recursive

Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ye Ting <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Long Qin <qin.long@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
4 years agoCryptoPkg/OpensslLib: Fix the documentation about submodule update
Long Qin [Thu, 12 Apr 2018 02:50:30 +0000 (10:50 +0800)]
CryptoPkg/OpensslLib: Fix the documentation about submodule update

This patch is to drop "--remote" option from the original suggested
submodule update command ("$ git submodule update --recursive
--remote") in HOWTO document.

"--remote" option will integrate changes from the upstream subproject
with the submodules's "current HEAD", instead of using the edk2
superproject's "recorded SHA-1".

It is important here for the edk2 consumers to updating the working
tree of the submodules to match the commit / release tag that the
superproject expects. So removing "--remote" option to fix this
documentation issue here.

Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ye Ting <ting.ye@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Long Qin <qin.long@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
4 years agoBaseTools: Fix one or more multiply defined symbols found issue
Feng, YunhuaX [Fri, 13 Apr 2018 05:20:31 +0000 (13:20 +0800)]
BaseTools: Fix one or more multiply defined symbols found issue

self.Guids update with package Guids will generate multiply defined
GUID symbols in AutoGen file

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yunhua Feng <yunhuax.feng@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoCryptoPkg/TlsLib: rewrite TlsSetCipherList()
Laszlo Ersek [Sat, 31 Mar 2018 15:33:14 +0000 (17:33 +0200)]
CryptoPkg/TlsLib: rewrite TlsSetCipherList()

Rewrite the TlsSetCipherList() function in order to fix the following
issues:

- Any cipher identifier in CipherId that is not recognized by
  TlsGetCipherMapping() will cause the function to return EFI_UNSUPPORTED.

  This is a problem because CipherId is an ordered preference list, and a
  caller should not get EFI_UNSUPPORTED just because it has an elaborate
  CipherId preference list. Instead, we can filter out cipher identifiers
  that we don't recognize, as long as we keep the relative order intact.

- CipherString is allocated on the stack, with 500 bytes.

  While processing a large CipherId preference list, this room may not be
  enough. Although no buffer overflow is possible, CipherString exhaustion
  can lead to a failed TLS connection, because any cipher names that don't
  fit on CipherString cannot be negotiated.

  Compute CipherStringSize first, and allocate CipherString dynamically.

- Finally, the "@STRENGTH" pseudo cipher name is appended to CipherString.
  (Assuming there is enough room left in CipherString.) This causes
  OpenSSL to sort the cipher list "in order of encryption algorithm key
  length".

  This is a bad idea. The caller specifically passes an ordered preference
  list in CipherId. Therefore TlsSetCipherList() must not ask OpenSSL to
  reorder the list, for any reason. Drop "@STRENGTH".

While at it, fix and unify the documentation of the CipherId parameter.

Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Qin Long <qin.long@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=915
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
4 years agoCryptoPkg/TlsLib: sanitize lib classes in internal header and INF
Laszlo Ersek [Sat, 31 Mar 2018 20:25:15 +0000 (22:25 +0200)]
CryptoPkg/TlsLib: sanitize lib classes in internal header and INF

"InternalTlsLib.h" includes "BaseCryptLib.h", but the lib class is not
listed in the INF file.

The INF file lists a good number of lib classes, but none of the lib class
headers are included by "InternalTlsLib.h".

Synchronize & sort both lists, while removing those library classes that
aren't actually needed. (IntrinsicLib and OpensslLib have no edk2 class
headers.)

Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Qin Long <qin.long@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=915
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
4 years agoCryptoPkg/TlsLib: pre-compute OpensslCipherLength in TlsCipherMappingTable
Laszlo Ersek [Sun, 1 Apr 2018 09:51:46 +0000 (11:51 +0200)]
CryptoPkg/TlsLib: pre-compute OpensslCipherLength in TlsCipherMappingTable

In the next patches, we'll need the lengths of the
TLS_CIPHER_MAPPING.OpensslCipher string fields. These lengths can be
computed at build time; add the new field "OpensslCipherLength", and
introduce the MAP() macro for populating it.

While at it, add some horizontal whitespace to "TlsCipherMappingTable",
and add a comma after the last element. This will come handy in a later
patch.

(The patch does not change the first two columns of
"TlsCipherMappingTable", which can be easily verified with "git show
--word-diff".)

Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Qin Long <qin.long@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=915
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
4 years agoCryptoPkg/TlsLib: use binary search in the TlsGetCipherMapping() function
Laszlo Ersek [Sat, 31 Mar 2018 15:06:39 +0000 (17:06 +0200)]
CryptoPkg/TlsLib: use binary search in the TlsGetCipherMapping() function

Improve the performance of the TlsGetCipherMapping() function by adopting
the binary search from DhcpFindOptionFormat()
[MdeModulePkg/Universal/Network/Dhcp4Dxe/Dhcp4Option.c].

Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Qin Long <qin.long@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=915
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
4 years agoCryptoPkg/TlsLib: replace TlsGetCipherString() with TlsGetCipherMapping()
Laszlo Ersek [Sun, 1 Apr 2018 10:53:23 +0000 (12:53 +0200)]
CryptoPkg/TlsLib: replace TlsGetCipherString() with TlsGetCipherMapping()

In the following patches it will be useful if the IANA CipherId lookup
returns a pointer to the whole matching IANA-to-OpenSSL mapping structure,
not just the OpenSSL cipher suite name. Rename TLS_CIPHER_PAIR and
TlsGetCipherString() to TLS_CIPHER_MAPPING and TlsGetCipherMapping()
respectively, and make the function return a pointer to
TLS_CIPHER_MAPPING.

Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Qin Long <qin.long@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Cc: Ting Ye <ting.ye@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=915
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
4 years agoNetworkPkg/TlsDxe: clean up byte order conversion for EfiTlsCipherList
Laszlo Ersek [Sat, 31 Mar 2018 14:36:39 +0000 (16:36 +0200)]
NetworkPkg/TlsDxe: clean up byte order conversion for EfiTlsCipherList

Fix the following style issues:

- "Data" is accessed through a pointer to UINT16 rather than to a pointer
  to EFI_TLS_CIPHER. While technically correct, UINT16 is harder to
  interpret against the UEFI spec.

- Array subscripting is written with weird *(Pointer + Offset)
  expressions, rather than with Pointer[Offset].

- The byte order is converted with HTONS(), while it should be NTOHS().
  Either way, use the Data1 and Data2 fields of EFI_TLS_CIPHER instead.

Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=915
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
4 years agoNetworkPkg/TlsDxe: verify DataSize for EfiTlsCipherList
Laszlo Ersek [Sat, 31 Mar 2018 14:04:10 +0000 (16:04 +0200)]
NetworkPkg/TlsDxe: verify DataSize for EfiTlsCipherList

TlsSetSessionData() shouldn't just ignore an incomplete EFI_TLS_CIPHER
element at the end of "Data":

- Generally speaking, malformed input for a security API is best rejected
  explicitly.

- Specifically speaking, the size of EFI_TLS_CIPHER is 2 bytes. If
  DataSize is 1 on input, then the initial check for (DataSize == 0) will
  fail, but then TlsSetCipherList() will be called with CipherNum=0.

Return EFI_INVALID_PARAMETER from TlsSetSessionData() if "Data" doesn't
contain a whole number of EFI_TLS_CIPHER elements. While at it, introduce
the dedicated variable CipherCount.

Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=915
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
4 years agoMdePkg/Include/Protocol/Tls.h: pack structures from the TLS RFC
Laszlo Ersek [Sat, 31 Mar 2018 14:20:37 +0000 (16:20 +0200)]
MdePkg/Include/Protocol/Tls.h: pack structures from the TLS RFC

The structures defined in RFC 5246 are not to have any padding between
fields or at the end; use the "pack" pragma as necessary.

Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=915
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Fu Siyuan <siyuan.fu@intel.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
4 years agoOvmfPkg/TlsAuthConfigLib: configure trusted cipher suites for HTTPS boot
Laszlo Ersek [Sat, 31 Mar 2018 23:27:43 +0000 (01:27 +0200)]
OvmfPkg/TlsAuthConfigLib: configure trusted cipher suites for HTTPS boot

Read the list of trusted cipher suites from fw_cfg and to store it to
EDKII_HTTP_TLS_CIPHER_LIST_VARIABLE.

The fw_cfg file will be formatted by the "update-crypto-policies" utility
on the host side, so that the host settings take effect in guest HTTPS
boot as well. QEMU forwards the file intact to the firmware. The contents
are forwarded by NetworkPkg/HttpDxe (in TlsConfigCipherList()) to
NetworkPkg/TlsDxe (TlsSetSessionData()) and TlsLib (TlsSetCipherList()).

Note: the development of the "update-crypto-policies" feature is underway
at this time. Meanwhile the following script can be used to generate the
binary file for fw_cfg:

  export LC_ALL=C
  openssl ciphers -V \
  | sed -r -n \
      -e 's/^ *0x([0-9A-F]{2}),0x([0-9A-F]{2}) - .*$/\\\\x\1 \\\\x\2/p' \
  | xargs -r -- printf -- '%b' > ciphers.bin

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Gary Ching-Pang Lin <glin@suse.com>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Gary Lin <glin@suse.com>
Tested-by: Gary Lin <glin@suse.com>
Reviewed-by: Long Qin <qin.long@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
[lersek@redhat.com: update commit msg and add script as requested by Gary]
[lersek@redhat.com: update commit msg as requested by Jiaxin]

4 years agoMdeModulePkg/PiSmmIpl: fix non-executable SMM RAM
Jian J Wang [Wed, 11 Apr 2018 06:12:12 +0000 (14:12 +0800)]
MdeModulePkg/PiSmmIpl: fix non-executable SMM RAM

This patch fixes an issue introduced by commit

  5b91bf82c67b586b9588cbe4bbffa1588f6b5926

and

  0c9f2cb10b7ddec56a3440e77219fd3ab1725e5c

This issue will only happen if PcdDxeNxMemoryProtectionPolicy is
enabled for reserved memory, which will mark SMM RAM as NX (non-
executable) during DXE core initialization. SMM IPL driver will
unset the NX attribute for SMM RAM to allow loading and running
SMM core/drivers.

But above commit will fail the unset operation of the NX attribute
due to a fact that SMM RAM has zero cache attribute (MRC code always
sets 0 attribute to reserved memory), which will cause GCD internal
method ConverToCpuArchAttributes() to return 0 attribute, which is
taken as invalid CPU paging attribute and skip the calling of
gCpu->SetMemoryAttributes().

The solution is to make use of existing functionality in PiSmmIpl
to make sure one cache attribute is set for SMM RAM. For performance
consideration, PiSmmIpl will always try to set SMM RAM to write-back.
But there's a hole in the code which will fail the setting write-back
attribute because of no corresponding cache capabilities. This patch
will add necessary cache capabilities before setting corresponding
attributes.

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>
Cc: Michael D Kinney <michael.d.kinney@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/SmmCore: add sanity check for SetMemoryAttributes
Jian J Wang [Wed, 11 Apr 2018 08:35:32 +0000 (16:35 +0800)]
MdeModulePkg/SmmCore: add sanity check for SetMemoryAttributes

Heap Guard feature needs enough memory and paging to work. Otherwise
calling SetMemoryAttributes to change page attribute will fail. This
patch add necessary check of result of calling SetMemoryAttributes.
This can help users to debug their problem in enabling this feature.

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: Star Zeng <star.zeng@intel.com>
4 years agoMdeModulePkg/DxeCore: add sanity check for SetMemoryAttributes
Jian J Wang [Wed, 11 Apr 2018 08:35:05 +0000 (16:35 +0800)]
MdeModulePkg/DxeCore: add sanity check for SetMemoryAttributes

Heap Guard feature needs enough memory and paging to work. Otherwise
calling SetMemoryAttributes to change page attribute will fail. This
patch add necessary check of result of calling SetMemoryAttributes.
This can help users to debug their problem in enabling this feature.

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: Star Zeng <star.zeng@intel.com>
4 years agoSignedCapsulePkg SystemCapsuleLib: Change some dbg level to DEBUG_INFO
Star Zeng [Wed, 28 Mar 2018 08:50:00 +0000 (16:50 +0800)]
SignedCapsulePkg SystemCapsuleLib: Change some dbg level to DEBUG_INFO

This debug message should be info instead of error. This patch is to
change the debug level to DEBUG_INFO.
DEBUG((DEBUG_ERROR, "checking FV....0x%08x - 0x%x\n",
       FvHeader, FvHeader->FvLength));    // "Mark"

This comment is inaccurate. This patch is to remove it.
//
// Check section
//

This debug message should be removed as FvHeader may have been out of
range FdStart and FdSize, and the loop will go to "Mark" above again if
FvHeader is not out of range FdStart and FdSize, and then that debug
message will be shown. This patch is to remove this debug message.
DEBUG((DEBUG_ERROR, "Next FV....0x%08x - 0x%x\n",
       FvHeader, FvHeader->FvLength));

Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Chasel Chiu <chasel.chiu@intel.com>
Cc: Amy Chan <amy.chan@intel.com>
Cc: Rangasai V Chaganty <rangasai.v.chaganty@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Chasel Chiu <chasel.chiu@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
4 years agoMdePkg DxeHstiLib: Fix ErrorString pointer incorrectly calculated
Star Zeng [Wed, 11 Apr 2018 06:18:38 +0000 (14:18 +0800)]
MdePkg DxeHstiLib: Fix ErrorString pointer incorrectly calculated

Fix ErrorString pointer is incorrectly calculated in
InternalHstiIsValidTable().

Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Liming Gao <liming.gao@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 agoArmVirtPkg/ArmVirtQemu: hook NvVarStoreFormattedLib into VariableRuntimeDxe
Laszlo Ersek [Wed, 11 Apr 2018 21:07:59 +0000 (23:07 +0200)]
ArmVirtPkg/ArmVirtQemu: hook NvVarStoreFormattedLib into VariableRuntimeDxe

In spite of both ArmVirtQemu and ArmVirtQemuKernel formatting the variable
store template at build time, link NvVarStoreFormattedLib into
VariableRuntimeDxe via NULL class resolution on both platforms. This lets
us test the depexes implemented in the previous patches.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
4 years agoArmVirtPkg/PlatformHasAcpiDtDxe: depend on gEfiVariableArchProtocolGuid
Laszlo Ersek [Wed, 11 Apr 2018 23:37:21 +0000 (01:37 +0200)]
ArmVirtPkg/PlatformHasAcpiDtDxe: depend on gEfiVariableArchProtocolGuid

PlatformHasAcpiDtDxe consumes the DynamicHii PCD called
"gArmVirtTokenSpaceGuid.PcdForceNoAcpi". The PcdGetBool() library call
terminates in gRT->GetVariable(), in the MdeModulePkg/Universal/PCD/Dxe
driver. Put "gEfiVariableArchProtocolGuid" on PlatformHasAcpiDtDxe's DEPEX
so that we not attempt the call before the PCD driver can successfully
read non-volatile variables.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.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: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPlatformPkg/PL031RealTimeClockLib: depend on gEfiCpuArchProtocolGuid
Laszlo Ersek [Thu, 12 Apr 2018 00:00:13 +0000 (02:00 +0200)]
ArmPlatformPkg/PL031RealTimeClockLib: depend on gEfiCpuArchProtocolGuid

The RealTimeClockLib class is declared under EmbeddedPkg, so that
platforms can provide the internals for the
EmbeddedPkg/RealTimeClockRuntimeDxe driver. In turn the driver produces
the Real Time Clock Arch Protocol, without which UEFI drivers cannot be
dispatched.

The PL031RealTimeClockLib instance calls gDS->SetMemorySpaceAttributes()
in the LibRtcInitialize() public function. This DXE service depends on the
CPU Arch Protocol. Add it to the depex.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Steve Capper <steve.capper@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPlatformPkg/NorFlashDxe: depend on gEfiCpuArchProtocolGuid
Laszlo Ersek [Wed, 11 Apr 2018 20:59:13 +0000 (22:59 +0200)]
ArmPlatformPkg/NorFlashDxe: depend on gEfiCpuArchProtocolGuid

NorFlashFvbInitialize() calls gDS->SetMemorySpaceAttributes() to mark the
varstore flash region as uncached. This DXE service depends on the CPU
Architectural protocol, and the DXE core is allowed to return
EFI_NOT_AVAILABLE_YET if it hasn't dispatched ArmPkg/Drivers/CpuDxe
earlier. Make the dependency explicit.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Steve Capper <steve.capper@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPlatformPkg/NorFlashDxe: cue the variable driver with NvVarStoreFormatted
Laszlo Ersek [Wed, 11 Apr 2018 20:50:18 +0000 (22:50 +0200)]
ArmPlatformPkg/NorFlashDxe: cue the variable driver with NvVarStoreFormatted

The BEFORE depex opcode that we currently use to force ourselves in front
of the variable driver cannot be combined with other depex opcodes.
Replace the depex with TRUE, and signal NvVarStoreFormattedLib through the
installation of "gEdkiiNvVarStoreFormattedGuid".

Platforms that rely on NorFlashDxe to format the variable store (as
opposed to formatting a variable store template through an FDF file, as
part of the build) should hook NvVarStoreFormattedLib into the variable
drivers they use, so that the latter await our cue.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Steve Capper <steve.capper@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPlatformPkg/NorFlashDxe: initialize varstore headers eagerly
Laszlo Ersek [Wed, 11 Apr 2018 20:18:24 +0000 (22:18 +0200)]
ArmPlatformPkg/NorFlashDxe: initialize varstore headers eagerly

The lazy initialization of the varstore FVB makes no longer sense at this
point:

- "mNorFlashInstanceTemplate.Initialize" is NULL;

- in NorFlashCreateInstance(), we only set Instance->Initialize to
  non-NULL -- namely NorFlashFvbInitialize() -- if the FVB stands for the
  variable store (see "ContainVariableStorage" / "SupportFvb");

- we call Instance->Initialize() from three places:

  - from NorFlashWriteSingleBlock(), which is too late for the variable
    read service ("variable write" depends on "variable read");

  - from InitializeFvAndVariableStoreHeaders(), but that is only reachable
    from NorFlashFvbInitialize(), i.e. recursively from
    Instance->Initialize() itself;

  - and from FvbRead(), which is never called by the variable driver, only
    by the FTW driver. However, the variable driver may read (not write)
    the memory-mapped varstore flash chip before the FTW driver is
    dispatched.

Therefore the lazy initialization is both superfluous and insufficient.
Initialize the varstore headers eagerly, before we install the FVB
protocol interface.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Steve Capper <steve.capper@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoEmbeddedPkg: introduce NvVarStoreFormattedLib
Laszlo Ersek [Wed, 11 Apr 2018 18:58:58 +0000 (20:58 +0200)]
EmbeddedPkg: introduce NvVarStoreFormattedLib

Some platforms don't format a variable store template at build time;
instead they format the non-volatile varstore flash chip during boot,
dynamically. Introduce NvVarStoreFormattedLib to enable such platforms to
delay the "variable read" service drivers until the platform specific
module(s) report that the variable store has been formatted.

The platform-specific module that performs the formatting during startup
is usually an FVB or MM FVB driver. Under the proposed scheme, it becomes
responsible for installing gEdkiiNvVarStoreFormattedGuid with a NULL
interface in the protocol database. In turn, the platform DSC will hook
NvVarStoreFormattedLib into the variable service driver, to make the
latter wait for the FVB driver. Platforms that need not delay the variable
service driver like this may still use the same FVB driver;
gEdkiiNvVarStoreFormattedGuid will simply be ignored.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Steve Capper <steve.capper@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPkg/CpuDxe: order CpuDxe after ArmGicDxe via protocol depex
Laszlo Ersek [Wed, 11 Apr 2018 16:52:25 +0000 (18:52 +0200)]
ArmPkg/CpuDxe: order CpuDxe after ArmGicDxe via protocol depex

Commit 61a7b0ec634f ("ArmPkg/Gic: force GIC driver to run before CPU arch
protocol driver", 2018-02-06) explains why CpuDxe should be dispatched
after ArmGicDxe.

To implement the ordering, we should use a regular protocol depex rather
than the less flexible AFTER opcode. ArmGicDxe installs
gHardwareInterruptProtocolGuid and gHardwareInterrupt2ProtocolGuid as one
of the last actions on its entry point stack; either of those is OK for
CpuDxe to wait for.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Steve Capper <steve.capper@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoArmPkg/ArmGicDxe: annotate protocol usage in "ArmGicDxe.inf"
Laszlo Ersek [Wed, 11 Apr 2018 16:40:49 +0000 (18:40 +0200)]
ArmPkg/ArmGicDxe: annotate protocol usage in "ArmGicDxe.inf"

"ArmGicDxe.inf" currently does not document how the protocols in the
[Protocols] section are used. Such comments help us analyze behavior, so
let's add them now.

- gHardwareInterruptProtocolGuid and gHardwareInterrupt2ProtocolGuid are
  always produced on the InterruptDxeInitialize() -> (GicV2DxeInitialize()
  | GicV3DxeInitialize()) -> InstallAndRegisterInterruptService() call
  path.

- gEfiCpuArchProtocolGuid is consumed in the CpuArchEventProtocolNotify()
  protocol notify callback. (Technically this is "conditional"; however
  the firmware cannot work without architectural protocols, so we can call
  it unconditional.)

While at it, drop the gArmGicDxeFileGuid comment from FILE_GUID; we're
going to make that GUID uninteresting soon.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Steve Capper <steve.capper@linaro.org>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoOmap35xxPkg/InterruptDxe: replace CPU Arch Protocol depex with notify
Laszlo Ersek [Wed, 11 Apr 2018 22:03:45 +0000 (00:03 +0200)]
Omap35xxPkg/InterruptDxe: replace CPU Arch Protocol depex with notify

In a later patch, we'll modify the depex of
"ArmPkg/Drivers/CpuDxe/CpuDxe.inf" (currently "AFTER gArmGicDxeFileGuid")
to "gHardwareInterruptProtocolGuid OR gHardwareInterrupt2ProtocolGuid".

Considering platforms that include "ArmPkg/Drivers/CpuDxe/CpuDxe.inf",
there are two classes:

(1) The platform gets its gHardwareInterruptProtocolGuid or
    gHardwareInterrupt2ProtocolGuid instance from
    "ArmPkg/Drivers/ArmGic/ArmGicDxe.inf". For such platforms, the
    upcoming CpuDxe change is not a problem, because commit 61a7b0ec634f
    made ArmGicDxe wait for the CPU Arch Protocol with a protocol notify.

(2) The platform gets its hardware interrupt protocol(s) from a different
    driver that has a hard depex on the CPU Arch Protocol. The upcoming
    CpuDxe change would lead to a loop in the DXE dispatch order.

In the edk2 tree, only "BeagleBoardPkg/BeagleBoardPkg.dsc" falls in class
(2), and the driver in question is "Omap35xxPkg/InterruptDxe". Port (most
of) commit 61a7b0ec634f to it.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Supreeth Venkatesh <Supreeth.Venkatesh@arm.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: Leif Lindholm <leif.lindholm@linaro.org>
4 years agoSignedCapsulePkg/SystemFirmwareReportDxe: Pass thru on same handle
Kinney, Michael D [Mon, 9 Apr 2018 22:47:19 +0000 (15:47 -0700)]
SignedCapsulePkg/SystemFirmwareReportDxe: Pass thru on same handle

https://bugzilla.tianocore.org/show_bug.cgi?id=928

Use HandleProtocol() to pass thru a SetImage() call to the
System FMP Protocol that must be on the same handle as the
FMP Protocol.

Cc: Jiewen Yao <jiewen.yao@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>
4 years agoSignedCapsulePkg/SystemFirmwareUpdateDxe: Single FMP
Kinney, Michael D [Sat, 31 Mar 2018 17:17:29 +0000 (10:17 -0700)]
SignedCapsulePkg/SystemFirmwareUpdateDxe: Single FMP

https://bugzilla.tianocore.org/show_bug.cgi?id=928

Uninstall all System FMP Protocols for the current FW device.

If an FMP Protocol for the current FW device is already present,
then install the new System FMP protocol onto the same handle as
the FMP Protocol.  Otherwise, install the FMP protocol onto a
new handle.

This supports use cases where multiple capsules for the
same system firmware device are processed on the same
boot of the platform.  It guarantees there is at most one
FMP protocol for each system firmware device.

Cc: Jiewen Yao <jiewen.yao@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>
4 years agoBaseTools: fix --genfds-multi-thread generate makefile issue
Feng, YunhuaX [Tue, 10 Apr 2018 01:12:49 +0000 (09:12 +0800)]
BaseTools: fix --genfds-multi-thread generate makefile issue

1. when inf file is binary module, not generate makefile,
  so need generate ffs with previous method.
2. generate Ui section maybe override and the string is not
  $(MODULE_NAME)
  like as:
  INF  RuleOverride = UI MdeModulePkg/Application/UiApp/UiApp.inf
3. Trim generate incorrect Offset.raw when some vfr not generate .lst
   file in Debug directory, Trim get the VFR name with the .c files
   replacement.
4. fix some depex file not generate issue

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yunhua Feng <yunhuax.feng@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: argument genfds-multi-thread create GenSec command issue
Feng, YunhuaX [Tue, 10 Apr 2018 01:10:41 +0000 (09:10 +0800)]
BaseTools: argument genfds-multi-thread create GenSec command issue

Issue:
  genfds-multi-thread create makefile before section file generation,
  so it get alignment is zero from empty file. It is incorrect.
solution:
  GenSec get section alignment from input file when the input alignment
  is zero.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yunhua Feng <yunhuax.feng@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Correct GenSec argument dummy free memory issue
Feng, YunhuaX [Tue, 10 Apr 2018 01:09:45 +0000 (09:09 +0800)]
BaseTools: Correct GenSec argument dummy free memory issue

Free DummyFileBuffer and set DummyFileBuffer to NULL.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yunhua Feng <yunhuax.feng@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Fix the build error caused by eca980c0c899
Yonghong Zhu [Tue, 10 Apr 2018 13:26:32 +0000 (21:26 +0800)]
BaseTools: Fix the build error caused by eca980c0c899

Roll back the fixed at build pcd collection to include the pcd in
Module and Library.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yonghong Zhu <yonghong.zhu@intel.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoOvmfPkg: remove BLOCK_MMIO_PROTOCOL and BlockMmioToBlockIoDxe
Laszlo Ersek [Tue, 10 Apr 2018 17:34:50 +0000 (19:34 +0200)]
OvmfPkg: remove BLOCK_MMIO_PROTOCOL and BlockMmioToBlockIoDxe

BLOCK_MMIO_PROTOCOL and BlockMmioToBlockIoDxe were introduced to OvmfPkg
in March 2010, in adjacent commits b0f5144676fa and efd82c5794ec. In the
past eight years, no driver or application seems to have materialized that
produced BLOCK_MMIO_PROTOCOL instances. Meanwhile the UEFI spec has
developed the EFI_RAM_DISK_PROTOCOL, which edk2 implements (and OVMF
includes) as RamDiskDxe.

Rather than fixing issues in the unused BlockMmioToBlockIoDxe driver,
remove the driver, together with the BLOCK_MMIO_PROTOCOL definition that
now becomes unused too.

Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Steven Shi <steven.shi@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=926
Reported-by: Steven Shi <steven.shi@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
4 years agoBaseTools: Remove unneeded files
Carsey, Jaben [Wed, 4 Apr 2018 15:01:45 +0000 (23:01 +0800)]
BaseTools: Remove unneeded files

These files are not used by any tool:
BaseTools/Source/Python/Common/DecClassObject.py
BaseTools/Source/Python/Common/DscClassObject.py
BaseTools/Source/Python/Common/FdfClassObject.py
BaseTools/Source/Python/Common/InfClassObject.py

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Remove EdkIIWorkspaceBuild.py from source code
Feng, YunhuaX [Wed, 4 Apr 2018 05:53:13 +0000 (13:53 +0800)]
BaseTools: Remove EdkIIWorkspaceBuild.py from source code

Remove EdkIIWorkspaceBuild.py from source code

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yunhua Feng <yunhuax.feng@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: copy a dictionary from InfClassObject to BuildReport
Carsey, Jaben [Wed, 4 Apr 2018 15:01:44 +0000 (23:01 +0800)]
BaseTools: copy a dictionary from InfClassObject to BuildReport

InfClassObject will be deleted.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Fix two cases that use GUID CName as PCD Value
Yonghong Zhu [Fri, 23 Mar 2018 03:39:39 +0000 (11:39 +0800)]
BaseTools: Fix two cases that use GUID CName as PCD Value

1. use CName format in components section:
  [Components]
   TestPkg/TestDriver.inf {
     <PcdsFixedAtBuild>
      PcdToken.PcdName |{GUID(TestGuid)}|VOID*|16
  }

2. Use Guid CName format in INF and the Guid is defined in the DEC
file but not write in driver's [Guids] section.
    PcdToken.PcdName  | {GUID(TestGuid)}

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: Fix the bug for VOID* pcd max size from component section
Yonghong Zhu [Sat, 17 Mar 2018 07:25:32 +0000 (15:25 +0800)]
BaseTools: Fix the bug for VOID* pcd max size from component section

When the Pcd defined in components section, its value's size is larger
than the value's size in [pcd] section, it cause build error, because
original code use the size get in [pcd] section as max size.

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: Fix size override issue for Void* Patchable pcd
Yonghong Zhu [Wed, 14 Mar 2018 12:20:23 +0000 (20:20 +0800)]
BaseTools: Fix size override issue for Void* Patchable pcd

when multiple driver link same library, and the drivers override the pcd
to different value in the DSC component section, it cause the pcd size
incorrect.

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: Fix a bug for Size incorrect of Void* Fixatbuild Pcd
Yonghong Zhu [Wed, 14 Mar 2018 08:51:04 +0000 (16:51 +0800)]
BaseTools: Fix a bug for Size incorrect of Void* Fixatbuild Pcd

when driver link library and there have pcd override in DSC component
section, in the library autogen file, the pcd's size is incorrect, the
size value is from DSC [pcd] section, but not from the override pcd
value that in the [component] section.

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: Fix a bug for VOID* type Fixatbuild Pcd in Library
Yonghong Zhu [Tue, 13 Mar 2018 14:44:01 +0000 (22:44 +0800)]
BaseTools: Fix a bug for VOID* type Fixatbuild Pcd in Library

The case is a FixedAtBuild VOID* PCD is used from a lib, but is set to a
different sized value in a module INF scope <PcdsFixedAtBuild> section.

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: Parse PCD GUID name in FILE statement issue
Yunhua Feng [Mon, 9 Apr 2018 01:07:09 +0000 (09:07 +0800)]
BaseTools: Parse PCD GUID name in FILE statement issue

FDF format as below:
FILE APPLICATION = PCD(PcdToken.PcdCName) {

}
when parse PCD, need get all PCDs from Platform and Packages,
use self.BuildObject[self.Platform, Arch] get some modules is wrong.
so use self.BuildObject[self.Platform, Arch, TargetName, ToolChainTag]
get all modules.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yunhua Feng <yunhuax.feng@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: refactor and remove out of date use of .keys()
Carsey, Jaben [Thu, 5 Apr 2018 23:14:02 +0000 (07:14 +0800)]
BaseTools: refactor and remove out of date use of .keys()

this is no longer required to make dictionary objects iterable.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: GenC - move content from both parts of if/else
Carsey, Jaben [Thu, 5 Apr 2018 23:14:01 +0000 (07:14 +0800)]
BaseTools: GenC - move content from both parts of if/else

move identical lines out of both if and else and move 1 level up.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: change more list to set
Carsey, Jaben [Thu, 5 Apr 2018 23:14:00 +0000 (07:14 +0800)]
BaseTools: change more list to set

potentially accelerate "in" testing
remove uncalled function

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: remove unneeded function call
Carsey, Jaben [Thu, 5 Apr 2018 23:13:59 +0000 (07:13 +0800)]
BaseTools: remove unneeded function call

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: change another list to set
Carsey, Jaben [Thu, 5 Apr 2018 23:13:58 +0000 (07:13 +0800)]
BaseTools: change another list to set

potentially accelerate "in" testing which is the use for this variable

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: optimize buildoptions loop
Carsey, Jaben [Thu, 5 Apr 2018 23:13:57 +0000 (07:13 +0800)]
BaseTools: optimize buildoptions loop

change a dict to a double defaultdict to prevent needing to seed innter values.
move "Value" determination under a conditional continue statement

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: simplify testing for existance and containing data
Carsey, Jaben [Thu, 5 Apr 2018 23:13:56 +0000 (07:13 +0800)]
BaseTools: simplify testing for existance and containing data

and remove a duplicate "if" block from 6 lines up.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: change list to set
Carsey, Jaben [Thu, 5 Apr 2018 23:13:55 +0000 (07:13 +0800)]
BaseTools: change list to set

Order is irelevant
duplication is auto-prevented

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: remove unused variables
Carsey, Jaben [Thu, 5 Apr 2018 23:13:54 +0000 (07:13 +0800)]
BaseTools: remove unused variables

some were populated, but never used after.
some were never used.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: replace a dict with a set
Carsey, Jaben [Thu, 5 Apr 2018 23:13:53 +0000 (07:13 +0800)]
BaseTools: replace a dict with a set

As we never use the values, just keep the keys in a set.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: sets are faster to check via "in" due to hashing
Carsey, Jaben [Thu, 5 Apr 2018 23:13:52 +0000 (07:13 +0800)]
BaseTools: sets are faster to check via "in" due to hashing

switch list to set:
1)we dont care about order
2)we only check for membership.

then remove ".keys()" from dict looping:
allow generators opportunity to optimize

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: defaultdict(set) allows us to just add to the set
Carsey, Jaben [Thu, 5 Apr 2018 23:13:51 +0000 (07:13 +0800)]
BaseTools: defaultdict(set) allows us to just add to the set

New sets will get created automatically when needed

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: remove uncalled functions
Carsey, Jaben [Thu, 5 Apr 2018 23:13:50 +0000 (07:13 +0800)]
BaseTools: remove uncalled functions

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: remove unused file
Carsey, Jaben [Thu, 5 Apr 2018 23:13:49 +0000 (07:13 +0800)]
BaseTools: remove unused file

ToolsDefClassObject didnt need Dictionary, it needed an import from there.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Pcds in [Components] are not display correct in the report
Yonghong Zhu [Mon, 2 Apr 2018 03:18:40 +0000 (11:18 +0800)]
BaseTools: Pcds in [Components] are not display correct in the report

The Pcd used in [Components] section, the PCD value is displayed
incorrect in the build report because the PCD default value was not
override.

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: Pcd not used info should not in Module PCD section
Yonghong Zhu [Mon, 2 Apr 2018 03:15:27 +0000 (11:15 +0800)]
BaseTools: Pcd not used info should not in Module PCD section

Pcds in Conditional Directives and Pcds not used are Platform Level
info, it should not display in Module PCD Section.

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: dont make temporary dict
Carsey, Jaben [Thu, 5 Apr 2018 14:00:24 +0000 (22:00 +0800)]
BaseTools: dont make temporary dict

just make the key list directly

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Remove FdfParserLite.py from source since it is not used
Feng, YunhuaX [Wed, 4 Apr 2018 05:53:13 +0000 (13:53 +0800)]
BaseTools: Remove FdfParserLite.py from source since it is not used

Remove FdfParserLite.py from source code since it is not used.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yunhua Feng <yunhuax.feng@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoMdeModulePkg/Gcd: Suppress incorrect compiler/analyzer warnings
Star Zeng [Sun, 8 Apr 2018 01:27:35 +0000 (09:27 +0800)]
MdeModulePkg/Gcd: Suppress incorrect compiler/analyzer warnings

It is caused by 0c9f2cb10b7ddec56a3440e77219fd3ab1725e5c
and false positive.
Initialize CpuArchAttributes to suppress incorrect
compiler/analyzer warnings.

Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@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 agoBaseTools: small cleanup
Carsey, Jaben [Wed, 4 Apr 2018 20:56:57 +0000 (04:56 +0800)]
BaseTools: small cleanup

just deleting else: then pass as they have no effect.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Autogen - change from list to set
Carsey, Jaben [Wed, 4 Apr 2018 20:56:56 +0000 (04:56 +0800)]
BaseTools: Autogen - change from list to set

by changing from list to set(), we can skip all the preprocessing
to prevent duplication and we dont need to convert to a set() later
on for each use

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: move RegEx to root of file and share it
Carsey, Jaben [Wed, 4 Apr 2018 20:56:55 +0000 (04:56 +0800)]
BaseTools: move RegEx to root of file and share it

make it easy to import and use by others

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: remove redundant check
Carsey, Jaben [Tue, 3 Apr 2018 22:34:06 +0000 (06:34 +0800)]
BaseTools: remove redundant check

The RegEx matches begining and end of the string, dont then check length.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: AutoGen - use the new shared RegEx
Carsey, Jaben [Tue, 3 Apr 2018 22:34:05 +0000 (06:34 +0800)]
BaseTools: AutoGen - use the new shared RegEx

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Add new RegEx pattern to GlobalData
Carsey, Jaben [Tue, 3 Apr 2018 22:34:04 +0000 (06:34 +0800)]
BaseTools: Add new RegEx pattern to GlobalData

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: GenFds - move RegEx compile
Carsey, Jaben [Tue, 3 Apr 2018 22:34:03 +0000 (06:34 +0800)]
BaseTools: GenFds - move RegEx compile

compile each RegEx once not in loops/functions

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Autogen - move RegEx compile
Carsey, Jaben [Tue, 3 Apr 2018 22:34:02 +0000 (06:34 +0800)]
BaseTools: Autogen - move RegEx compile

compile each RegEx once not in loops/functions

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Remove unused code from Misc
Carsey, Jaben [Tue, 3 Apr 2018 21:03:10 +0000 (05:03 +0800)]
BaseTools: Remove unused code from Misc

remove the functions and classes
remove any imports of these

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Workspace - use built in OrderedDict instead of custom version.
Carsey, Jaben [Tue, 3 Apr 2018 21:03:09 +0000 (05:03 +0800)]
BaseTools: Workspace - use built in OrderedDict instead of custom version.

We dont use any feature added by custom dictionary class.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: use combined version of OrderedDict
Carsey, Jaben [Tue, 3 Apr 2018 21:03:08 +0000 (05:03 +0800)]
BaseTools: use combined version of OrderedDict

since we need order and a default entry, use collections dicts to
auto generate.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: use built in OrderedDict instead of custom version.
Carsey, Jaben [Tue, 3 Apr 2018 21:03:07 +0000 (05:03 +0800)]
BaseTools: use built in OrderedDict instead of custom version.

We dont use any feature added by custom dictionary class.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Eot remove unused code
Carsey, Jaben [Tue, 3 Apr 2018 21:03:06 +0000 (05:03 +0800)]
BaseTools: Eot remove unused code

2 functions and a dictionary that are not used.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools - AutoGen - replace custom dictionary class with python standard one
Carsey, Jaben [Tue, 3 Apr 2018 21:03:05 +0000 (05:03 +0800)]
BaseTools - AutoGen - replace custom dictionary class with python standard one

We have a custom ordered dictionary class.  works fine with python OrderedDict version.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: remove unused import statement
Carsey, Jaben [Tue, 3 Apr 2018 21:03:04 +0000 (05:03 +0800)]
BaseTools: remove unused import statement

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Eot tool never populates this dictionary
Carsey, Jaben [Tue, 3 Apr 2018 21:03:03 +0000 (05:03 +0800)]
BaseTools: Eot tool never populates this dictionary

we initialize this dict and then check it's contents, but never add items.
we can remove it without any effect.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: use built in dict instead of custom version.
Carsey, Jaben [Tue, 3 Apr 2018 21:03:02 +0000 (05:03 +0800)]
BaseTools: use built in dict instead of custom version.

We dont use any feature added by custom dictionary class.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoBaseTools: Use local variable for list of constants.
Carsey, Jaben [Tue, 3 Apr 2018 21:03:01 +0000 (05:03 +0800)]
BaseTools: Use local variable for list of constants.

instead of listing in multiple places, use a single list.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Yonghong Zhu <yonghong.zhu@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jaben Carsey <jaben.carsey@intel.com>
Reviewed-by: Yonghong Zhu <yonghong.zhu@intel.com>
4 years agoMdeModulePkg/Gcd: Filter gCpu->SetMemoryAttributes() calls
Kinney, Michael D [Mon, 2 Apr 2018 22:40:24 +0000 (06:40 +0800)]
MdeModulePkg/Gcd: Filter gCpu->SetMemoryAttributes() calls

This patch fixes an issue with VlvTbltDevicePkg introduced
by commit 5b91bf82c67b586b9588cbe4bbffa1588f6b5926.

The history is as below.
To support heap guard feature, 14dde9e903bb9a719ebb8f3381da72b19509bc36
added support for SetMemorySpaceAttributes() to handle page attributes,
but after that, a combination of CPU arch attributes and other attributes
was not allowed anymore, for example, UC + RUNTIME. It is a regression.
Then 5b91bf82c67b586b9588cbe4bbffa1588f6b5926 was to fix the regression,
and we thought 0 CPU arch attributes may be used to clear CPU arch
attributes, so 0 CPU arch attributes was allowed to be sent to
gCpu->SetMemoryAttributes().

But some implementation of CPU driver may return error for 0 CPU arch
attributes. That fails the case that caller just calls
SetMemorySpaceAttributes() with none CPU arch attributes (for example,
RUNTIME), and the purpose of the case is not to clear CPU arch attributes.

This patch filters the call to gCpu->SetMemoryAttributes()
if the requested attributes is 0.  It also removes the #define
INVALID_CPU_ARCH_ATTRIBUTES that is no longer used.

Cc: Heyi Guo <heyi.guo@linaro.org>
Cc: Yi Li <phoenix.liyi@huawei.com>
Cc: Renhao Liang <liangrenhao@huawei.com>
Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Signed-off-by: Star Zeng <star.zeng@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Michael D Kinney <michael.d.kinney@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: use mnemonics for FXSAVE(64)/FXRSTOR(64)
Laszlo Ersek [Fri, 23 Mar 2018 19:54:19 +0000 (20:54 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: use mnemonics for FXSAVE(64)/FXRSTOR(64)

NASM introduced FXSAVE / FXRSTOR support in commit 900fa5b26b8f ("NASM
0.98p3-hpa", 2002-04-30), which commit stands for the nasm-0.98p3-hpa
release.

NASM introduced FXSAVE64 / FXRSTOR64 support in commit 3a014348ca15
("insns: add FXSAVE64/FXRSTOR64, drop np prefix", 2010-07-07), which was
part of the "nasm-2.09" release.

Edk2 requires nasm-2.10 or later for use with the GCC toolchain family,
and nasm-2.12.01 or later for use with all other toolchain families.
Replace the binary encoding of the FXSAVE(64)/FXRSTOR(64) instructions
with mnemonics.

I verified that the "Ia32/SmiException.obj", "X64/SmiEntry.obj" and
"X64/SmiException.obj" files are rebuilt after this patch, without any
change in content.

This patch removes the last instructions encoded with DBs from
PiSmmCpuDxeSmm.

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: remove DBs from SmmRelocationSemaphoreComplete32()
Laszlo Ersek [Fri, 2 Feb 2018 04:36:22 +0000 (05:36 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: remove DBs from SmmRelocationSemaphoreComplete32()

(1) SmmRelocationSemaphoreComplete32() runs in 32-bit mode, so wrap it in
    a (BITS 32 ... BITS 64) bracket.

(2) SmmRelocationSemaphoreComplete32() currently compiles to:

000002AE  C6050000000001    mov byte [dword 0x0],0x1
000002B5  FF2500000000      jmp dword [dword 0x0]

    where the first instruction is patched with the contents of
    "mRebasedFlag" (so that (*mRebasedFlag) is set to 1), and the second
    instruction is patched with the address of
    "mSmmRelocationOriginalAddress" (so that we jump to
    "mSmmRelocationOriginalAddress").

    In its current form the first instruction could not be patched with
    PatchInstructionX86(), given that the operand to patch is not encoded
    in the trailing bytes of the instruction. Therefore, adopt an
    EAX-based version, inspired by both the IA32 and X64 variants of
    SmmRelocationSemaphoreComplete():

000002AE  50                push eax
000002AF  B800000000        mov eax,0x0
000002B4  C60001            mov byte [eax],0x1
000002B7  58                pop eax
000002B8  FF2500000000      jmp dword [dword 0x0]

    Here both instructions can be patched with PatchInstructionX86(), and
    the DBs can be replaced with native NASM syntax.

(3) Turn the "mRebasedFlagAddr32" and "mSmmRelocationOriginalAddressPtr32"
    variables into markers that suit PatchInstructionX86().

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmmInitStack" with PatchInstructionX86()
Laszlo Ersek [Fri, 2 Feb 2018 03:46:26 +0000 (04:46 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmmInitStack" with PatchInstructionX86()

Rename the variable to "gPatchSmmInitStack" so that its association with
PatchInstructionX86() is clear from the declaration, change its type to
X86_ASSEMBLY_PATCH_LABEL, and patch it with PatchInstructionX86(). This
lets us remove the binary (DB) encoding of some instructions in
"SmmInit.nasm".

The size of the patched source operand is (sizeof (UINTN)).

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: eliminate "gSmmJmpAddr" and related DBs
Laszlo Ersek [Fri, 2 Feb 2018 03:12:51 +0000 (04:12 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: eliminate "gSmmJmpAddr" and related DBs

The IA32 version of "SmmInit.nasm" does not need "gSmmJmpAddr" at all (its
PiSmmCpuSmmInitFixupAddress() variant doesn't do anything either). We can
simply use the NASM syntax for the following Mixed-Size Jump:

> jmp PROTECT_MODE_CS : dword @32bit

The generated object code for the instruction is unchanged:

00000182  66EA5A0000000800  jmp dword 0x8:0x5a

(The NASM manual explains that putting the DWORD prefix after the colon
":" reflects the intent better, since it is the offset that is a DWORD.
Thus, that's what I used. However, both syntaxes are interchangeable,
hence the ndisasm output.)

The X64 version of "SmmInit.nasm" appears to require "gSmmJmpAddr";
however that's accidental, not inherent:

- Bring LONG_MODE_CODE_SEGMENT from
  "UefiCpuPkg/PiSmmCpuDxeSmm/PiSmmCpuDxeSmm.h" to "SmmInit.nasm" as
  LONG_MODE_CS, same as PROTECT_MODE_CODE_SEGMENT was brought to the IA32
  version as PROTECT_MODE_CS earlier.

- Apply the NASM-native Mixed-Size Jump syntax again, but jump to the
  fixed zero offset in LONG_MODE_CS. This will produce no relocation
  record at all. Add a label after the instruction.

- Modify PiSmmCpuSmmInitFixupAddress() to patch the jump target backwards
  from the label. Because we modify the DWORD offset with a DWORD access,
  the segment selector is unharmed in the instruction, and we need not set
  it from PiCpuSmmEntry().

According to "objdump --reloc", the X64 version undergoes only the
following relocations, after this patch:

> RELOCATION RECORDS FOR [.text]:
> OFFSET           TYPE              VALUE
0000000000000095 R_X86_64_PC32     SmmInitHandler-0x0000000000000004
00000000000000e0 R_X86_64_PC32     mRebasedFlag-0x0000000000000004
00000000000000ea R_X86_64_PC32     mSmmRelocationOriginalAddress-0x0000000000000004

Therefore the patch does not regress
<https://bugzilla.tianocore.org/show_bug.cgi?id=849> ("Enable XCODE5 tool
chain for UefiCpuPkg with nasm source code").

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmmCr0" with PatchInstructionX86()
Laszlo Ersek [Fri, 2 Feb 2018 01:10:05 +0000 (02:10 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmmCr0" with PatchInstructionX86()

Like "gSmmCr4" in the previous patch, "gSmmCr0" is not only used for
machine code patching, but also as a means to communicate the initial CR0
value from SmmRelocateBases() to InitSmmS3ResumeState(). In other words,
the last four bytes of the "mov eax, Cr0Value" instruction's binary
representation are utilized as normal data too.

In order to get rid of the DB for "mov eax, Cr0Value", we have to split
both roles, patching and data flow. Introduce the "mSmmCr0" global (SMRAM)
variable for the data flow purpose. Rename the "gSmmCr0" variable to
"gPatchSmmCr0" so that its association with PatchInstructionX86() is clear
from the declaration, change its type to X86_ASSEMBLY_PATCH_LABEL, and
patch it with PatchInstructionX86(), to the value now contained in
"mSmmCr0".

This lets us remove the binary (DB) encoding of "mov eax, Cr0Value" in
"SmmInit.nasm".

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmmCr4" with PatchInstructionX86()
Laszlo Ersek [Fri, 2 Feb 2018 01:10:05 +0000 (02:10 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmmCr4" with PatchInstructionX86()

Unlike "gSmmCr3" in the previous patch, "gSmmCr4" is not only used for
machine code patching, but also as a means to communicate the initial CR4
value from SmmRelocateBases() to InitSmmS3ResumeState(). In other words,
the last four bytes of the "mov eax, Cr4Value" instruction's binary
representation are utilized as normal data too.

In order to get rid of the DB for "mov eax, Cr4Value", we have to split
both roles, patching and data flow. Introduce the "mSmmCr4" global (SMRAM)
variable for the data flow purpose. Rename the "gSmmCr4" variable to
"gPatchSmmCr4" so that its association with PatchInstructionX86() is clear
from the declaration, change its type to X86_ASSEMBLY_PATCH_LABEL, and
patch it with PatchInstructionX86(), to the value now contained in
"mSmmCr4".

This lets us remove the binary (DB) encoding of "mov eax, Cr4Value" in
"SmmInit.nasm".

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmmCr3" with PatchInstructionX86()
Laszlo Ersek [Fri, 2 Feb 2018 00:48:56 +0000 (01:48 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmmCr3" with PatchInstructionX86()

Rename the variable to "gPatchSmmCr3" so that its association with
PatchInstructionX86() is clear from the declaration, change its type to
X86_ASSEMBLY_PATCH_LABEL, and patch it with PatchInstructionX86(). This
lets us remove the binary (DB) encoding of some instructions in
"SmmInit.nasm".

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: remove unneeded DBs from X64 SmmStartup()
Laszlo Ersek [Fri, 2 Feb 2018 00:23:17 +0000 (01:23 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: remove unneeded DBs from X64 SmmStartup()

(This patch is the 64-bit variant of commit e75ee97224e5,
"UefiCpuPkg/PiSmmCpuDxeSmm: remove unneeded DBs from IA32 SmmStartup()",
2018-01-31.)

The SmmStartup() function executes in SMM, which is very similar to real
mode. Add "BITS 16" before it and "BITS 64" after it (just before the
@LongMode label).

Remove the manual 0x66 operand-size override prefixes, for selecting
32-bit operands -- the sizes of our operands trigger NASM to insert the
prefixes automatically in almost every spot. The one place where we have
to add it back manually is the LGDT instruction. In the LGDT instruction
we also replace the binary 0x2E prefix with the normal NASM syntax for CS
segment override.

The stores to the Control Registers were always 32-bit wide; the source
code only used RAX as source operand because it generated the expected
object code (with NASM compiling the source as if in BITS 64). With BITS
16 added, we can use the actual register width in the source operands
(EAX).

This patch causes NASM to generate byte-identical object code (determined
by disassembling both the pre-patch and post-patch versions, and comparing
the listings), except:

> @@ -231,7 +231,7 @@
>  000001D2  6689D3            mov ebx,edx
>  000001D5  66B800000000      mov eax,0x0
>  000001DB  0F22D8            mov cr3,eax
> -000001DE  662E670F0155F6    o32 lgdt [cs:ebp-0xa]
> +000001DE  2E66670F0155F6    o32 lgdt [cs:ebp-0xa]
>  000001E5  66B800000000      mov eax,0x0
>  000001EB  80CC02            or ah,0x2
>  000001EE  0F22E0            mov cr4,eax

The only difference is the prefix list order, it changes from:

- 0x66, 0x2E, 0x67

to

- 0x2E, 0x66, 0x67

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: patch "XdSupported" with PatchInstructionX86()
Laszlo Ersek [Thu, 1 Feb 2018 23:17:13 +0000 (00:17 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: patch "XdSupported" with PatchInstructionX86()

"mXdSupported" is a global BOOLEAN variable, initialized to TRUE. The
CheckFeatureSupported() function is executed on all processors (not
concurrently though), called from SmmInitHandler(). If XD support is found
to be missing on any CPU, then "mXdSupported" is set to FALSE, and further
processors omit the check. Afterwards, "mXdSupported" is read by several
assembly and C code locations.

The tricky part is *where* "mXdSupported" is allocated (defined):

- Before commit 717fb60443fb ("UefiCpuPkg/PiSmmCpuDxeSmm: Add paging
  protection.", 2016-11-17), it used to be a normal global variable,
  defined (allocated) in "SmmProfile.c".

- With said commit, we moved the definition (allocation) of "mXdSupported"
  into "SmiEntry.nasm". The variable was defined over the last byte of a
  "mov al, 1" instruction, so that setting it to FALSE in
  CheckFeatureSupported() would patch the instruction to "mov al, 0". The
  subsequent conditional jump would change behavior, plus all further read
  references to "mXdSupported" (in C and assembly code) would read back
  the source (imm8) operand of the patched MOV instruction as data.

  This trick required that the MOV instruction be encoded with DB.

In order to get rid of the DB, we have to split both roles: we need a
label for the code patching, and "mXdSupported" has to be defined
(allocated) independently of the code patching. Of course, their values
must always remain in sync.

(1) Reinstate the "mXdSupported" definition and initialization in
    "SmmProfile.c" from before commit 717fb60443fb. Change the assembly
    language definition ("global") to a declaration ("extern").

(2) Define the "gPatchXdSupported" label (type X86_ASSEMBLY_PATCH_LABEL)
    in "SmiEntry.nasm", and add the C-language declaration to
    "SmmProfileInternal.h". Replace the DB with the MOV mnemonic (keeping
    the imm8 source operand with value 1).

(3) In CheckFeatureSupported(), whenever "mXdSupported" is set to FALSE,
    patch the assembly code in sync, with PatchInstructionX86().

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmiCr3" with PatchInstructionX86()
Laszlo Ersek [Thu, 1 Feb 2018 22:40:29 +0000 (23:40 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmiCr3" with PatchInstructionX86()

Rename the variable to "gPatchSmiCr3" so that its association with
PatchInstructionX86() is clear from the declaration, change its type to
X86_ASSEMBLY_PATCH_LABEL, and patch it with PatchInstructionX86(). This
lets us remove the binary (DB) encoding of some instructions in
"SmiEntry.nasm".

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmiStack" with PatchInstructionX86()
Laszlo Ersek [Thu, 1 Feb 2018 22:23:59 +0000 (23:23 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmiStack" with PatchInstructionX86()

Rename the variable to "gPatchSmiStack" so that its association with
PatchInstructionX86() is clear from the declaration. Also change its type
to X86_ASSEMBLY_PATCH_LABEL.

Unlike "gSmbase" in the previous patch, "gSmiStack"'s patched value is
also de-referenced by C code (in other words, it is read back after
patching): the InstallSmiHandler() function stores "CpuIndex" to the given
CPU's SMI stack through "gSmiStack". Introduce the local variable
"CpuSmiStack" in InstallSmiHandler() for calculating the stack location
separately, then use this variable for both patching into the assembly
code, and for storing "CpuIndex" through it.

It's assumed that "volatile" stood in the declaration of "gSmiStack"
because we used to read "gSmiStack" back for de-referencing; with that use
gone, we can remove "volatile" too. (Note that the *target* of the pointer
was never volatile-qualified.)

Finally, replace the binary (DB) encoding of "mov esp, imm32" in
"SmiEntry.nasm".

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmbase" with PatchInstructionX86()
Laszlo Ersek [Thu, 1 Feb 2018 22:01:08 +0000 (23:01 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: patch "gSmbase" with PatchInstructionX86()

Rename the variable to "gPatchSmbase" so that its association with
PatchInstructionX86() is clear from the declaration, change its type to
X86_ASSEMBLY_PATCH_LABEL, and patch it with PatchInstructionX86(). This
lets us remove the binary (DB) encoding of some instructions in
"SmiEntry.nasm".

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg/PiSmmCpuDxeSmm: remove *.S and *.asm assembly files
Laszlo Ersek [Thu, 1 Feb 2018 21:35:18 +0000 (22:35 +0100)]
UefiCpuPkg/PiSmmCpuDxeSmm: remove *.S and *.asm assembly files

All edk2 toolchains use NASM for compiling X86 assembly source code. We
plan to remove X86 *.S and *.asm files globally, in order to reduce
maintenance and confusion:

http://mid.mail-archive.com/4A89E2EF3DFEDB4C8BFDE51014F606A14E1B9F76@SHSMSX104.ccr.corp.intel.com
https://lists.01.org/pipermail/edk2-devel/2018-March/022690.html
https://bugzilla.tianocore.org/show_bug.cgi?id=881

Let's start with UefiCpuPkg/PiSmmCpuDxeSmm: remove the *.S and *.asm
dialects (both Ia32 and X64) of the SmmInit, SmiEntry, SmiException and
MpFuncs sources.

Cc: Eric Dong <eric.dong@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Andrew Fish <afish@apple.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdePkg/BaseLib: add PatchInstructionX86()
Laszlo Ersek [Thu, 1 Feb 2018 21:00:40 +0000 (22:00 +0100)]
MdePkg/BaseLib: add PatchInstructionX86()

Some edk2 modules generate X86 machine code at module execution time by:

- compiling "template" code with NASM at module build time,

- linking the object code into the module,

- and patching the immediate (constant) operands of some instructions when
  the module is executed.

Add a helper function to BaseLib so that the C code performing the
patching is easier to read and maintain.

The implementation in this patch is taken mainly from Mike Kinney's
mailing list messages at
<http://mid.mail-archive.com/E92EE9817A31E24EB0585FDF735412F5B895C360@ORSMSX113.amr.corp.intel.com>,
<http://mid.mail-archive.com/E92EE9817A31E24EB0585FDF735412F5B898BF66@ORSMSX112.amr.corp.intel.com>.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoMdePkg/BaseLib.h: state preprocessing conditions in comments after #endifs
Laszlo Ersek [Thu, 1 Feb 2018 20:08:28 +0000 (21:08 +0100)]
MdePkg/BaseLib.h: state preprocessing conditions in comments after #endifs

"#endif" preprocessing directives near the top of "BaseLib.h" helpfully
repeat the preprocessing conditions from their matching "#if", "#ifdef",
and "#ifndef" directives. This practice has been less followed recently;
supplement the missing comments.

Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=866
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
4 years agoUefiCpuPkg PiSmmCpuDxeSmm: Refine some comments about SmmMemoryAttribute
Star Zeng [Wed, 28 Feb 2018 05:21:59 +0000 (13:21 +0800)]
UefiCpuPkg PiSmmCpuDxeSmm: Refine some comments about SmmMemoryAttribute

1. Fix some "support" to "supported".
2. Fix some "set" to "clear" in ClearMemoryAttributes interface.
3. Remove redundant comments for GetMemoryAttributes interface.

Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
4 years agoMdeModulePkg SmmMemoryAttribute.h: Refine some comments
Star Zeng [Wed, 28 Feb 2018 05:20:34 +0000 (13:20 +0800)]
MdeModulePkg SmmMemoryAttribute.h: Refine some comments

1. Fix some "support" to "supported".
2. Fix some "set" to "clear" in ClearMemoryAttributes interface.
3. Remove redundant comments for GetMemoryAttributes interface.

Cc: Jian J Wang <jian.j.wang@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: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>