> On Nov 6, 2015, at 1:51 AM, Zeng, Star wrote:
>
> On 2015/10/23 14:04, Andrew Fish wrote:
>>
>>> On Oct 22, 2015, at 10:10 PM, Zeng, Star wrote:
>>>
>>> On 2015/10/22 23:29, Andrew Fish wrote:
>>>>
>>>>> On Oct 21,
aligned to make it easier for the OS to map.
Thanks,
Andrew Fish
> Thanks,
> Shubha
> Shubha D. ramanishubharam...@gmail.com
> shubharam...@yahoo.com
> ___
> edk2-devel mailing list
> edk2-dev
Please review.
To build with Xcode without installing NASM I had to make the following change.
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
Please review the following patches as they are required to build with Xcode
6.3.2.
Thanks,
Andrew Fish
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
03-MdeModulePkg-Variable-add-mising-VA_COPY.patch
I attached again here:
Thanks,
Andrew Fish
> From: edk2-devel [mailto:edk2-devel-boun...@lists.01.org] On Behalf Of Andrew
> Fish
> Sent: Tuesday, November 10, 2015 3:18 PM
> To: edk2-devel
> Cc: Tian, Feng
> Subject: [edk2] Md
> On Nov 10, 2015, at 7:04 AM, Andrew Fish wrote:
>
>
>> On Nov 10, 2015, at 12:06 AM, Gao, Liming wrote:
>>
>> Andrew:
>> The patch is missing.
>>
>
> Liming,
>
> My sent mail shows these 3 patches attached?
> 0001-MdeModule
c Mon Sep 17 00:00:00 2001
From: andrew fish
Date: Mon, 9 Nov 2015 22:46:32 -0800
Subject: [PATCH 4/4] MdePkg: BaseLib don't require NASM with Xcode
Don't require NASM for simple edk2 Xcode projects.
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Andre
> On Nov 10, 2015, at 7:04 AM, Andrew Fish wrote:
>
>
>> On Nov 10, 2015, at 12:06 AM, Gao, Liming wrote:
>>
>> Andrew:
>> The patch is missing.
>>
>
> Liming,
>
> My sent mail shows these 3 patches attached?
> 0001-MdeModule
> On Nov 10, 2015, at 7:04 AM, Andrew Fish wrote:
>
>>
>> On Nov 10, 2015, at 12:06 AM, Gao, Liming > <mailto:liming@intel.com>> wrote:
>>
>> Andrew:
>> The patch is missing.
>>
>
> Liming,
>
> My sent mail shows thes
TOD_DEFINES_BUILDRULEORDER =
'BUILDRULEORDER'
I realize now that my previous test may have been flawed since I had NASM
installed.
Thanks,
Andrew Fish
> -Jordan
>
> On 2015-11-10 18:30:39, Gao, Liming wrote:
>> This patch is committed at r18766.
>>
>> Th
> On Nov 10, 2015, at 8:06 PM, Jordan Justen wrote:
>
> On 2015-11-10 19:18:01, Andrew Fish wrote:
>> On Nov 10, 2015, at 7:11 PM, Jordan Justen
>> wrote:
>> Liming, Andrew,
>>
>> Wasn't BUILDRULEORDER supposed to help with this?
>>
&g
; index 000..7f6cbd9
> --- /dev/null
> +++ b/MdeModulePkg/Include/Protocol/FileExplorer.h
> @@ -0,0 +1,75 @@
> +/** @file
> +
> + This file explorer protocol defines defines a set of interfaces for
> + how to do file explorer.
> +
> +Copyright (c) 2006 - 2010, I
"do the correct thing". So I think the objection
>is the shell was written to strictly conform to the latest UEFI specification,
>when it could have been made backward compatible with a range of UEFI
>specification versions.
Thanks,
Andrew Fish
>
> Thanks
> Samer
>
TPL of your event from making forward progress.
gBS->Stall() does not yield, it is no different than running code.
So there is only one context, you can print it out any time you want.
Thanks,
Andrew Fish
PS When folks yell at us for not having threads in EFI we point them at:
https:/
k. Every thing
else is an event that runs in the main EFI context.
https://github.com/tianocore/edk2/blob/master/MdeModulePkg/Core/Dxe/Event/Tpl.c#L126
Thanks,
Andrew Fish
> On Thu, Nov 12, 2015 at 9:27 AM, Andrew Fish wrote:
>
>>
>>> On Nov 11, 2015, at 11:51 PM, Mi
> On Nov 12, 2015, at 3:22 AM, Andrew Fish wrote:
>
>
>> On Nov 12, 2015, at 12:49 AM, Michael Zimmermann
>> wrote:
>>
>> Stall was just an example, I can also use DEBUG. My timer interval is
>> 100ms(I converted it from the 100ns unit).
>>
&
> On Nov 12, 2015, at 3:28 AM, Andrew Fish wrote:
>
>>
>> On Nov 12, 2015, at 3:22 AM, Andrew Fish wrote:
>>
>>
>>> On Nov 12, 2015, at 12:49 AM, Michael Zimmermann
>>> wrote:
>>>
>>> Stall was just an example, I can also
sume it is your event and figure out if you are doing something slow at an
elevated TPL, calling a service at too hight a TPL., or have TPL related
deadlock based on the TPL restrictions.
Thanks,
Andrew Fish
PS Sorry if this is more detail than you needed, but I figured other folks
might find
from the build output directory
and see what code it maps to.
The X86 version of all this lives:
https://github.com/tianocore/edk2/blob/master/UefiCpuPkg/Library/CpuExceptionHandlerLib/CpuExceptionCommon.c
Thanks,
Andrew Fish
> On Thu, Nov 12, 2015 at 7:39 PM, Andrew Fish wrote:
>
>&
FI/OS code that calls gRT->SetVariable().
For the edk2 the format is defined here:
https://github.com/tianocore/edk2/blob/master/MdeModulePkg/Include/Guid/VariableFormat.h
This is the DXE Driver that reads and writes the variables.
https://github.com/tianocore/edk2/tree/master/MdeModulePkg/Un
ecause they're not thread safe.
>
Is it one of the edk2 timer libs? If so which one?
Thanks,
Andrew Fish
> The reason is that a hw counter gets reset and started for this and if this
> happens from an interrupt(UART uses delay functions) while the thread mode
> is whiting fo
libs? If so which one?
>
> Nope it's my own, platform specific one(my device hasn't been merged to the
> official sources yet).
>
Most of the other TimerLib implementations use a free running counter.
Thanks,
Andrew Fish
> Michael
>
>
> On Sat, Nov 14, 201
> On Nov 17, 2015, at 5:31 PM, winddy wrote:
>
> Hi RuiYu,
> patch file is attached.
I don't see the attachment. I've had trouble attaching patches to the mailing
list too?
Thanks,
Andrew Fish
> Thanks.
>
>
>
> On Nov 17, 2015, at 5:49 PM, winddy wrote:
>
> Hi Fish,
> I see sent mail has attached file.
> The content of the patch is as below:
>
The mail I got from the mailing list did not include this content?
Thanks,
Andrew Fish
>
> --- MdeMod
timization flags for VC++ are more complex than /O1, as it is uses a very
specific /O1 variant and also uses other flags to turn on link time
optimization.
Thanks,
Andrew Fish
> How do I also add /O1 (that's capitol letter "O" 1).
> Thanks,
> Shub
ce will ASSERT on any calls.
You want to use the instances of the library that call the PPI/Protocol.
https://github.com/tianocore/edk2/tree/master/MdePkg/Library/PeiPcdLib
https://github.com/tianocore/edk2/tree/master/MdePkg/Library/DxePcdLib
Thanks,
Andrew Fish
> __
k at a log of the build. So add
-y build.log (or --report-file=build.log). It will show what library instances
got picked for the driver and what PCD values that the driver is using.
Thanks,
Andrew Fish
> Thanks for your help,
>
> ---
75b/BaseTools/Source/Python/Workspace/WorkspaceDatabase.py",
>> line 2814, in __init__
>> [16:50:59]W: [Step 2/5] self.Conn.execute("PRAGMA synchronous=OFF")
>> [16:50:59]W: [Step 2/5] DatabaseError: database disk image is malformed
Try removing Conf/.c
Is there a special linker setting I am missing or a technique?
>> I am using the latest UEFI code from
>> https://github.com/tianocore/edk2.git
>>
>
> I am sorry, but since I don't have access to DS-5, I am not sure how
> to debug
ing) to bit-rot.
>
> I think this issue is not directly related to OVMF.
>
The strange thing about this is the failure is loading the DXE Core from the
DXE IPL, so all of PEI has already run and been dispatched?
#2 0x000000822f0e in PeiLoadImageLoadImage at
/home/miki/vcs/git/edk2/M
or Mach-O we ended up linking the image with the offset of the PE/COFF (0x240)
so the tool that converts to PE/COFF does not have to apply relocations.
Thanks,
Andrew Fish
> Reviewed-by: Eugene Cohen
>
> Eugene
>
>> -Original Message-
>> From: Ard Biesheuvel
LowData), // %0
"=d" (HighData) // %1
: "c" (Index) // %2
);
return (((UINT64)HighData) << 32) | LowData;
}
Not sure why the Vlv2TbltDevicePkg needed to add that assembly code? I would
also point out that the inline as
t of the bitcode, so when the native code generation happens it
is very easy for the compiler to optimize.
Thanks,
Andrew Fish
~/work/Compiler>cat bit.c
#define IN
typedef unsigned int UINT32;
typedef unsigned long long UINT64;
UINT64
AsmReadMsr64 (
IN UINT32I
ou will need to remove your
Conf/boots_def.txt file to get the BaseTools/Conf/tools_def.template file
copied over by the edksetup script.
Thanks,
Andrew Fish
> ___
> edk2-devel mailing list
> edk2-devel@lists.01.org
> https://lists.01.org/
__TEXT,__cstring,cstring_literals
L_.str1:## @.str1
.space 1
.section__DATA,__data
.globl _Two ## @Two
.align 3
_Two:
.quad L_.str1
.comm _Three,8,3 ## @Three
.subs
k2 C include files. This should prevent the
name mangling of the edk2 C code functions. I think folks have ported C++ code
in the past, so I guess it is possible.
>From a quick search of the internet I see that those relocations are for a
>different memory model so it seems like the compiler
nux C++ runtime is not going to run properly on
Windows.
Thanks,
Andrew Fish
> -Ravi
>> Date: Mon, 14 Dec 2015 19:37:50 +0100
>> Subject: Re: [edk2] GenFW Error 3000 Invalid WriteRelocations64() errors
>> From: ard.biesheu...@linaro.org
>> To: mvrravika...@live.com
>
BaseTools maintainer know of any thing that would prevent this from
working, and would it be possible to store this information in the Conf/
directory to make parallel builds possible?
Thanks,
Andrew Fish
___
edk2-devel mailing list
edk2-devel@lists.01
;
I would also point out that Subsystem ID and Subsystem Vendor ID can be used
for this type of binding too. The Subsystem values exist to deal with cases
when the same chip is used on different vendors PCI plug in cards.
Thanks,
Andrew Fish
> Mike
>
>
>> -Original Message-
CTION syntax like:
FILE Type $(NAMED_GUID) [Options] FileName
SECTION GUIDED $(GUID_CNAME) [auth] FILENAME
Thanks,
Andrew Fish
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
just using a FILE
statement.
I was trying to do something like this, but use a GUID_CNAME.
FILE APPLICATION = PCD(gEfiIntelFrameworkModulePkgTokenSpaceGuid.PcdShellFile) {
SECTION PE32 = EdkShellBinPkg/MinimumShell/$(EDK_DXE_ARCHITECTURE)/Shell.efi
}
Thanks,
Andrew Fish
> -Origin
"%d %d %d %d %d %d\n"
.subsections_via_symbols
So as long as your bit code is getting the correct call prefix it means the bug
is in the code generator. If it happens a lot you may be able to figure it from
a simple command line var arg example. Just start with all your compiler flag
there is a linker flag, -object_path_lto
$(DEST_DIR_DEBUG)/$(BASE_NAME).lto, that needs to get added to store temp
output of the LTO generation to support debug, but there is not a compiler flag
required. The Xcode clang flags are very GCC like, but the linker is totally
GCCLTO53: 101KB
> 46KB
> CLANGLTO38:107KB
> 48KB
Xcode104K 49K
> GCC49:184KB
7; (x86_64).
(lldb) dis -m -b -n main
a.out`main
a.out`main:
a.out[0x10f98] <+0>: 55pushq %rbp
a.out[0x10f99] <+1>: 48 89 e5 movq %rsp, %rbp
a.out[0x10f9c] <+4>: 31 c0 xorl %eax, %eax
a.out[0x10f9e] <+6>: 5dpopq %rbp
a.out[0x10f9f
It is my understanding
that the Xcode linker just links the bit code like normal (pulls in the
symbols that are needed). Then this linked bit code blob is sent to an LLVM
dynamic library to do the code gen. Maybe different version of LLVM are used by
the different linkers in your case, or
pre-UEFI Spec RISC-V UEFI system.
Thanks,
Andrew Fish
> Good discussion.
>
>> -Original Message-
>> From: edk2-devel [mailto:edk2-devel-boun...@lists.01.org] On Behalf Of
>> Jordan Justen
>> Sent: Monday, May 9, 2016 10:46 AM
>> To: Yao, Jiewen ; C
)
{
EFI_HOB_GUID_TYPE *GuidHob;
GUID SavedDataGuid = PRE_PI_EXTRACT_GUIDED_SECTION_DATA_GUID;
I would guess if you make SavedDataGuid static, or global the issue will go
away.
Thanks,
Andrew Fish
>
> Please help what I am missi
ork/Compiler>cat memcpy.S | grep memcpy
So I think this is a bug in the code, it is just being hidden by the optimizers
on x86, and the CompilerIntrinsicsLib on ARM. If this code was compiled with
clang NOOPT it would likely also fail to link.
Thanks,
Andrew Fish
> On Thu, May 12, 2016 a
rg function that was NOT EFIAPI it might fail?
Try my example with your fix (__builtin_ms_va_start) but remove
__attribute__((ms_abi)) and see if you get the correct answer.
I'll ping our compiler team to see if there is a better answer.
Thanks,
Andrew Fish
> In my current edk2 llvm b
but -target
has the potential to give you others. I'm not sure there is a way to dump the
targets in clang.
You could try: -target x86_64-pc-win32-elf vs -arch x86_64. I'm guessing that
is what gets used for MinGW?
Thanks,
Andrew Fish
> Does Mac support qemu?
nd of
thing. Maybe the answer is don't do it :).
Also as I mentioned in my other mail there may be a cross compiler target in
clang that can produce EFIAPI natively.
Thanks,
Andrew Fish
> This is similar to the rule that all functions implemented in assembly must
> be
> EFIA
> On May 12, 2016, at 9:32 AM, Andrew Fish wrote:
>
>
>> On May 12, 2016, at 8:49 AM, Kinney, Michael D
>> wrote:
>>
>> Andrew,
>>
>> I recall an issues with var args in the past (not CLANG related) that
>> requires the
>> U
is why the standard committee
decided that the significant performance wins were worth the unexpected result
for "reasonable" type casts. It is worth pointing out that Java gets the
benefits of type-based optimizations without these drawbacks because it doesn't
have unsafe po
in UEFI spec 2.7. This works?
I was being paranoid about the version, as some one could merge the wrong thing
and mess it up. It is a public repo so it can be cloned by others. If the Cpu
DXE driver published the version protocol then it takes a git commit to remove
it and it is less likely to g
small Python script to generate the 8-4-4-4-12 and EFI_GUID C
structure format at the same time.
Thanks,
Andrew Fish
> --
> Thanks,
> David Van Arnem
> ___
> edk2-devel mailing list
> edk2-devel@lists.01.org
lumes/Case/Janus/Vendor/X/XPkg Path=StdLib/StdLib.inc
getWs() /Volumes/Case/edk2-git/Vendor/X/XPkg
For a test case you can shift to an edk2/ and then try to include StdLib.inc.
!include StdLib/StdLib.inc
Thanks,
Andrew Fish
___
edk2-devel mailing lis
uid[11], guid[12], guid[13], guid[14], guid[15])
Contributed-under: TianoCore Contribution Agreement 1.0
Thanks,
Andrew Fish
> Michael
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
ready-to-use string produced by str(guid) which matches my manual output
> exactly.
>
> Source: https://docs.python.org/2/library/uuid.html#uuid.UUID.bytes
>
Sorry I was using UUID.bytes_le, but I see you are using UUID.bytes that
explains the difference.
Thanks,
Andrew Fish
>
s issue?
>
> Root
> -->edk2-BaseTools-win32
> -->edk2
> -->Conf
> -->Build
>
The DSC file that was including, that failed, was also in an alternate path.
So try moving things in parts. Move the Nt32 to one PACKAGE_PATH and the StdLib
to another.
Thanks,
Andrew
join(self.AlterRoot, self.File)
elif self.Root:
RealFile = os.path.join(self.Root, self.File)
if len (mws.getPkgPath()) == 0:
return FILE_NOT_FOUND, os.path.join(self.AlterRoot, RealFile)
else:
return FILE_NOT_FOUND, "%s i
> On May 16, 2016, at 6:55 AM, Andrew Fish wrote:
>
>
>> On May 15, 2016, at 11:48 PM, Gao, Liming wrote:
>>
>> PACKAGES_PATH
>
> It seems to depend on sequence? I'm using master non Windows so running the
> Python directly. I just moved StdLi
> On May 16, 2016, at 9:56 AM, Andrew Fish wrote:
>
>>
>> On May 16, 2016, at 6:55 AM, Andrew Fish wrote:
>>
>>
>>> On May 15, 2016, at 11:48 PM, Gao, Liming wrote:
>>>
>>> PACKAGES_PATH
>>
>> It seems to depend on sequ
point EfiBootServicesData is basically freed
back for use by the OS.
The Shell vs. BDS probably just change the address that gets allocated and
moves it a range that shows the conflict.
Thanks,
Andrew Fish
___
edk2-devel mailing list
edk2-devel@l
use it?
>
Don't know about Linux stuff. What address did AllocatePages() return and what
size was it? 0x00F42000@0x does not seem to contain that info?
Thanks,
Andrew Fish
> console=ttyS0,115200n8 loglevel=8 rootwait root=/dev/mtdblock0
> mtdparts=RAM0:0x0
7;ve found is:
>
> http://lists.llvm.org/pipermail/llvm-dev/2016-January/094012.html
>
Seems reasonable to follow John's advice. The llvm team is the most aggressive
in terms of optimizing out undefined behavior, so they will end up setting the
defacto standards i
st be returned as EfiReservedMemoryType. The OS may use addresses of memory
ranges that are not described in the memory map at its own discretion.
Thanks,
Andrew Fish
>
>
> -Original Message-
> From: Laszlo Ersek [mailto:ler...@redhat.com]
> Sent: Tuesday, May 17, 20
rack down that logic.
Thanks,
Andrew Fish
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
This error message is for a single path in PACKAGES_PATH, so I think the error
message could be more clear.
For example my bug using ; vs : on Unix gets an error message that looks like.
: error 000E: PACKAGES_PATH doesn't exist
/Volumes/Case/Janus/edk2/;/Volumes/Case/Janus/Vendor/;/V
rk
in your case?
Thanks,
Andrew Fish
> Thanks
> Liming
> -Original Message-
> From: edk2-devel [mailto:edk2-devel-boun...@lists.01.org] On Behalf Of Andrew
> Fish
> Sent: Wednesday, May 18, 2016 7:24 AM
> To: edk2-devel
> Subject: [edk2] [BaseTools] RuleOverrid
connect all and starts all the EFI drivers.
When you boot directly from the BDS only drivers that are needed for boot are
started. So maybe the issue is the EFI driver does something that the OS driver
depends on?
Thanks,
Andrew Fish
> Please kindly help.
>
>
> Thanks,
> Meenakshi
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
Reviewed-by: Andrew Fish
> On May 17, 2016, at 5:49 AM, Pedroa wrote:
>
> Add 'Alt' key map.
>
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Pedroa Liu
> ---
> EmulatorPkg/Unix/Host/X11GraphicsWindow.c | 16 +++
data out from a program was a recommendation from the LLVM
compiler team, as there is not a truly portable way to figure out what the
compiler is going to do.
Maybe someone has a better idea?
Thanks,
Andrew Fish
> Some examples of this additional flexibility are:
>
> UINT8 TRU
> On May 18, 2016, at 2:27 PM, Kinney, Michael D
> wrote:
>
> Hi Andrew,
>
> Stay tuned. I have an additional [RFC] I am writing up now for associating
> C structures with PCDs.
>
I look forward to reviewing the new RFC.
Thanks,
Andrew Fish
> These value f
Depex means all the above architectural protocols are available and all of the
EFI services will just work. Some times people think no Depex is like a Depex
of TRUE. The reality is the Depex exist to solve all the chicken and egg
initializations problems in getting CPU/Chipse
\R9Tip\edk2\BaseTools\Source\Python\GenFds\Section.py.
>
Thanks, I'll take a look as I'm seeing applications that don't have .bin
extensions getting included too.
Thanks,
Andrew Fish
> Thanks
> Liming
> From: edk2-devel [mailto:edk2-devel-boun...@lists.01.org] On Be
vs AST. The real value in this scheme is unifying
the include infrastructure for C, ASL, and PCD. C can already access C
structure members, seems like reinventing the wheel to come up with a complex
scheme to abstract how to access C structure elements when a C pointer solves
that problem.
"The real art is knowing what to leave out, not what to put in" quote from
Steve Jobs.
Thanks,
Andrew Fish
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
> On May 19, 2016, at 12:16 PM, Kinney, Michael D
> wrote:
>
> Andrew,
>
> My responses embedded below.
>
Mike,
Likewise.
> Mike
>
>> -Original Message-
>> From: edk2-devel [mailto:edk2-devel-boun...@lists.01.org] On Behalf Of
>> A
> On May 18, 2016, at 5:44 PM, Andrew Fish wrote:
>
>>
>> On May 18, 2016, at 5:43 PM, Gao, Liming wrote:
>>
>> Andrew:
>> |.bin will search the file with the postfix .bin in INF file directory and
>> its output directory. So, there may be more tha
> On May 19, 2016, at 8:18 PM, Gao, Liming wrote:
>
> Thanks Andrew. I will continue to investigate it.
>
>>> F = 'FileNoExtention'
>>> os.path.splitext(F)[1]
''
>>> '' in '.out'
True
So I think the
Reviewed-by: Andrew FIsh
Thanks you for the quick fix.
Andrew Fish
> On May 20, 2016, at 1:42 AM, Liming Gao wrote:
>
> GenFds GenSection will search the output file based on the file extension.
> If the output file has no extension, it should be skip.
>
> Cc: Andrew F
you are trying to change some
kind of fundamental platform policy.
Thanks,
Andrew Fish
> Thanks
> Iru
>
> On Thu, May 19, 2016 at 3:22 PM, Gao, Liming wrote:
>
>> You can create ProtocolNotification function for gEfiCcidProtocolGuid in
>> ccidboot driver. If so,
T, Leaf.leaf),
SIZE_OF(STRUCT, Leaf.leaf));
return 0;
}
So maybe the complexity tradeoff is how you support bitfields (union with a
real type and granularity of access is that type) and how pedantic the type
checking (C is not type safe, so this is no different than a cast in the C
code) is
## 0x3c
.space 3
.globl _Testc ## @Testc
.align 2
_Testc:
.byte 192 ## 0xc0
.space 3
OK I think I'm convinced now it is likely we will be able to support bitfields.
Thanks,
Andrew Fish
PS Mike if you lo
ctory does not exist and no
files are found. It looks like the issue is __GetEFIOutPutPath__() as it is
using self.InfFileName and that has not been adjusted for PACKAGES_PATH?
Thanks,
Andrew Fish
> On May 20, 2016, at 1:34 AM, Gao, Liming wrote:
>
> Andrew:
> Thanks for your poi
you get the EFI Spec structure packing.
Thanks,
Andrew Fish
diff --git a/BaseTools/BinWrappers/PosixLike/StructuredPcdAutoGen.py
b/BaseTools/BinWrappers/PosixLike/StructuredPcdAutoGen.py
new file mode 100755
index 000..82a9b3b
--- /dev/null
+++ b/BaseTools/BinWrappers/PosixLike
> On May 22, 2016, at 3:04 PM, Andrew Fish wrote:
>
> Mike,
>
> I did a quick prototype that just outputs the current PCD DSC syntax with an
> extra | C typedef name at the end when you build. The example includes a
> structure with bit-fields and a device path.
about your advocated scheme.
Thanks,
Andrew Fish
> Your requires the entire structure value to be set every time. There is
> no concept of sparsely overriding fields. Also, if a backwards compatible
> change is made to a structure (i.e. converting reserved fields to new fields),
>
on like this for files:
Build/X/DEBUG_XCODE5/X64/MdeModulePkg/Universal/Acpi/AcpiTableDxe/AcpiTableDxe/OUTPUT/
Thanks,
Andrew Fish
> I try Ovmf platform. Its AcpiTables can be found and be built into FFS.
>
> Thanks
> Liming <>
> <>From: af...@apple.com <
spent a couple of hours on the prototype on Sunday,
and it supports structures, bit fields, and device paths. Since this scheme
leverages the existing C infrastructure it should currently support any
possible C construct that could be used.
Thanks,
Andrew Fish
> Thanks
> Liming <
erable via an
API.
> DSC File
> =
> [PcdsFixedAtBuild.common]
> gMdeModulePkgTokenGuid.PcdFakeDevicePathExample|DEVICE_PATH(UsbClass(0x,0x,3,1,1))
>
> If a PCD is a single node (no end node), we could have a 2nd operator
> called DEVICE_PATH_NODE()
mon.UEFI_APPLICATION,[LibraryClasses.Common.UEFI_DRIVER]
You could also add a DSC file to SpacelabsPkg and just build the things you car
about.
Thanks,
Andrew Fish
> [Packages]
> MdePkg/MdePkg.dec
> MdeModulePkg/MdeModulePkg.dec
> StdLib/StdLib.dec
>
> [LibraryCl
odulePkg/Universal/Acpi/S3SaveStateDxe/S3SaveStateDxe.inf
INF
MdeModulePkg/Universal/Acpi/BootScriptExecutorDxe/BootScriptExecutorDxe.inf
Thanks,
Andrew Fish
> On May 23, 2016, at 8:34 AM, Andrew Fish wrote:
>
>
>> On May 23, 2016, at 5:52 AM, Gao, Liming wrote:
>>
>>
ld output directory, I'd like to fix the Build
directories to be the actual path in the tree. I guess I have to fix both the
DSC and FDF derived paths to match. So there are two locations I need to path.
I'm willing to modify our copy of the tools to get back our old behavior.
> On May 24, 2016, at 7:44 AM, Andrew Fish wrote:
>
>
>> On May 24, 2016, at 6:29 AM, Gao, Liming wrote:
>>
>> Andrew:
>> If the source module INF is listed in FDF file, it is also required to be
>> placed into DSC file.
>
> Well they are the
g: cast to 'void *' from smaller integer type 'UINTN' (aka
'unsigned int')
[-Wint-to-void-pointer-cast]
return (void *) (UINTN) Memory;
^
1 warning generated.
~/work/Compiler>clang -arch i386 -S void.c
Thanks,
Andrew Fish
___
edk2-devel mailing list
edk2-devel@lists.01.org
https://lists.01.org/mailman/listinfo/edk2-devel
ore about cross compilers. For example
*XCODE5_X64_CC_FLAGS sets -target x86_64-pc-win32-macho to compile using the
Windows/EFI Calling convention but to generate a Mach-O (Image format on a Mac
like ELF or PE/COFF) to make the debugger happy.
I did not see IA32 or X64 targets for CLANG35 in m
=1
> jmpq *.LJTI3_0(,%rcx,8)
> … …
>
> In below Simics, you can clearly see the “jmp qword ptr 0xfffcdd54[rcx*8]”
> instruction cause the processor access logic address 0xfffcdd54 (not
> we hoped 0xfffcdd54)
Steven,
Can you dump the b
= PcdGet16 (PcdPlatformBootTimeOut);
>> +
>> + Black.Blue = Black.Green = Black.Red = Black.Reserved = 0;
>> + White.Blue = White.Green = White.Red = White.Reserved = 0xFF;
>
> I know this came from Nt32, but how about making these global vars?
>
Do you need to
ding style (6.8 C Function Layout,
>> "Initializing a variable as part of its declaration is illegal" --
>> although it clearly doesn't cover variables with static storage duration.)
>>
>
> I do see that in the older style guide, but not the newer ones (2.0+).
>
1 - 100 of 764 matches
Mail list logo
