vanjeff [Mon, 26 Mar 2007 07:00:30 +0000 (07:00 +0000)]
1. Removed definition for EFI_1_1 from module ConSplitter
2. Removed old Intel Framework specification extension for Media Firmware file path from Uefi/UefiSpec.h and UefiLib/UefiNotTiano.c
3. Added EFI revision check in UefiLib/UefiNotTiano.c and removed one EFI_SPECIFICATION_VERSION wrapper in Uefi/UefiSpec.h.
qwang12 [Thu, 22 Mar 2007 10:36:04 +0000 (10:36 +0000)]
Add return EFI_INVALID_PARAMETER if pointer type parameter for UEFI and Tiano Decompress protocol function is NULL. The check is necessary. The protocols is built based on library instance of the Decompress Library. The Library class in MDE library spec 0.60e only define to do ASSERT for NULL pointer input.
vanjeff [Thu, 22 Mar 2007 03:03:32 +0000 (03:03 +0000)]
1. Removed AsmSwitchStackAndBackingStore from BaseLib.h
2. Modified SwitchStack definition update to MWG 0.60e.
3. Added BaseLib/Ipf/InternalSwitchStack.c and removed BaseLib/Ipf/Non-existing.c
4. Updated any references of AsmSwitchStackAndBackingStore to SwichStack on IPF.
vanjeff [Wed, 21 Mar 2007 01:27:01 +0000 (01:27 +0000)]
1. Removed #ifdef EDK_RELEASE_VERSION from all c files for all modules
2. Removed #ifdef EFI_SPECIFICATION_VERSION from all c files for all modules
3. Added comments for file VariableWorker.c
qwang12 [Tue, 20 Mar 2007 10:04:54 +0000 (10:04 +0000)]
Change the TPL lock level to EFI_TPL_NOTIFY. Variable Service is called in ResetSystem runtime service. As there is no restriction stated in UEFI spec about TPL restriction for ResetSystem, so SCT does test it under NOTIFY level. This fix can be temperary. We may revert it back after discussion with Architects..
vanjeff [Tue, 20 Mar 2007 02:44:09 +0000 (02:44 +0000)]
1. Removed #ifdef SNP_DEBUG and used debug lib to output information
2. Rename PcdPlatformNoCsmSupport to PcdPlatformCsmSupport.
3 Removed #ifdef EFI_NO_CSM and used PcdPlatformCsmSupport.
4. Removed #ifdef ALIGNMENT from C Variable.c and added 2 files Variable/Pei/Ipf/VariableWorker.c and Variable/Pei/VariableWorker.c
qwang12 [Mon, 19 Mar 2007 09:05:41 +0000 (09:05 +0000)]
Change the EfiAquireLock with RaiseTPL as the call to Boot Service UnloadImage can be nested in another call to UnloadImage. So RaiseTPL is proper (EfiAcquireLock will ASSERT when these this API are nested in the call chain) in this situation to server the sychronization purposes.
The same change has been done to Boot Service Exit with the same reason applied.
klu2 [Fri, 16 Mar 2007 09:12:20 +0000 (09:12 +0000)]
Fix the issue that the variable Unicode string may be not aligned on 16-bit boundary in flash (it dependent on ALIGNMENT macro in EdkModulePkg\Universal\Variable\Pei\x64\VarMachine.h). The unalignment Unicode string will lead assert of StrCmp, StrCopy, StrLen function in Mde library.
vanjeff [Thu, 15 Mar 2007 02:33:16 +0000 (02:33 +0000)]
1. Remove #ifdef _MSC_EXTENSION_ from all source files
2. Remove #ifdef _GCC_ from all resource files
3. Remove #if 0 from all resource files
4. Remove #ifdef MDE_CPU_IPF and #ifdef MDE_CPU_EBC
5. Remove #progam from Decode.c
6. Remove #ifdef TIANO_EXTENSION_FLAG from source files
7. Remove #ifdef SUPPORT_IPV6, since code base doesn't support IPV6 at all.
8. Remove #ifdef SpecialNowaitVersion from pxe_bc_mtftp.c, since it's only used for test.
vprabhal [Wed, 14 Mar 2007 23:45:03 +0000 (23:45 +0000)]
changed the guid assignment from constant array to hi and low UINT64s (Intel compiler was using a global variable to assign an array which is not OK at run time)
vprabhal [Wed, 14 Mar 2007 23:44:44 +0000 (23:44 +0000)]
changed the guid assignment from constant array to hi and low UINT64s (Intel compiler was using a global variable to assign an array which is not OK at run time)
vprabhal [Wed, 14 Mar 2007 23:43:34 +0000 (23:43 +0000)]
changed the guid assignment from constant array to hi and low UINT64s (Intel compiler was using a global variable to assign an array which is not OK at run time)
qwang12 [Mon, 12 Mar 2007 09:58:28 +0000 (09:58 +0000)]
Add a lock to protect the critical region in Service APIs for gEfiBlockIoProtocolGuid and gEfiSimpleFileSystemProtocolGuid Protocol to prevent re-entrance of the same API from from different TPL level. In UEFI 2.1 spec, it is state that the service API for this Protocol is callable at EFI_TPL_CALLBACK level.
qwang12 [Mon, 12 Mar 2007 09:48:53 +0000 (09:48 +0000)]
Add a lock to protect the critical region in Service APIs for gEfiBlockIoProtocolGuid and gEfiSimpleFileSystemProtocolGuid Protocol to prevent re-entrance of the same API from from different TPL level. In UEFI 2.1 spec, it is state that the service API for this Protocol is callable at EFI_TPL_CALLBACK level.
qwang12 [Mon, 12 Mar 2007 09:42:34 +0000 (09:42 +0000)]
Add a lock to protect the critical region in Service APIs for gEfiSimpleTextOutProtocolGuid Protocol to prevent re-entrance of the same API from from different TPL level. In UEFI 2.1 spec, it is state that the service API for this Protocol is callable at EFI_TPL_NOTIFY level.
qwang12 [Mon, 12 Mar 2007 07:56:16 +0000 (07:56 +0000)]
Add a lock to protect the critical region in Service APIs for UEFI Runtime Variable Service to prevent re-entrance of the variable service API from from different allowable TPL level. In UEFI 2.x and EFI 1.10 spec, EFI_TPL_CALLBACK is the allowable TPL level for Variable services.
qwang12 [Mon, 12 Mar 2007 07:47:33 +0000 (07:47 +0000)]
Add a lock to protect the critical region in Service APIs for gEfiBlockIoProtocolGuid Protocol to prevent re-entrance of the same API from from different TPL level.
jwang36 [Mon, 12 Mar 2007 06:57:33 +0000 (06:57 +0000)]
Fixed EDKT532 and EDKT533
- Support GUID CName in <Events>, <SystemTables>, <DataHubs>, <HiiPackages>, <Hobs> in msa file
- Support auto-insert related header file of DebugLib and UefiBootServicesTableLib according to module type
jwang36 [Sat, 10 Mar 2007 14:23:39 +0000 (14:23 +0000)]
Commented out the pre-compile option of IPF ARCH for ICC tool chain. There's an issue in it which will cause incremental build failure. And it seems that there's not much improvement on build performance of Intel compiler. Comment out the pre-compile functionality for further investigation.
jwang36 [Thu, 8 Mar 2007 01:44:17 +0000 (01:44 +0000)]
- Fixed PVCS tracker 484 by merging the GenDll and GenEfi macro in BuildMacro.xml to GenEfi, moving out GenDll and GenEfi macro calling inside macro EFI_SECTION_TE and EFI_SECTION_PE32, and generating the calling of GenEfi macro according to the module type (binary or library).
- Cleaned the coding style of AutoGen.java
lgao4 [Mon, 5 Mar 2007 06:25:35 +0000 (06:25 +0000)]
Correct APPLICATION_EFI ffs key file type from EFI_FV_FILETYPE_DRIVER to EFI_FV_FILETYPE_APPLICATION, and remove depex section from application ffs layout.
klu2 [Fri, 2 Mar 2007 08:21:43 +0000 (08:21 +0000)]
In IdeBus driver block I/O read/write interface, it will always try to use UDMA mode first, if fails, then try to use PIO mode.
This may cause side effects if disk does not support UDMA mode and also let bad performance.
The modification is to record supported mode and only try supported mode when do disk read/write.
Also this patch combines the code for read/readex/write/writex to reduce code size.
qhuang8 [Wed, 28 Feb 2007 07:41:28 +0000 (07:41 +0000)]
1. Fix an issue about the calculation of GlyphBufferSize
2. Fix a bug in handling the overlapped case by reverse-copying the destine region & source region
3. Fix an issue that the leading spaces could not be displayed in highlight mode
xli24 [Wed, 14 Feb 2007 07:19:58 +0000 (07:19 +0000)]
Temporarily disable /WX for HelloWorld module, for there is a bug in tool, which causes inconsistency between UefiApplicationLib and AutoGen.c leading to warning of ICC compiler.