Re: [edk2-devel] [PATCH] CryptoPkg/FltUsedLib: Add FltUsedLib for float.

["Guomin Jiang" <guomin.jiang@intel.com>]

I guess the key point is /GL option. IntrinsicLib omit the /GL option to avoid the build error, but it abandon the optimization meanwhile.
I will do a test: create a simplest application and just depend IntrinsicLib and add /GL option to compare the different with and without /GL.

> -----Original Message-----
> From: Ni, Ray <ray.ni@intel.com>
> Sent: Thursday, April 23, 2020 11:32 AM
> To: Jiang, Guomin <guomin.jiang@intel.com>; devel@edk2.groups.io; Kinney,
> Michael D <michael.d.kinney@intel.com>; ard.biesheuvel@linaro.org
> Cc: lersek@redhat.com; macarl@microsoft.com
> Subject: RE: [edk2-devel] [PATCH] CryptoPkg/FltUsedLib: Add FltUsedLib for
> float.
> 
> Guomin,
> Can you investigate why moving _fltused from CryptoPkg/IntrinsicLib to
> MdePkg/BaseLib saves size?
> 
> Thanks,
> Ray
> 
> > -----Original Message-----
> > From: Jiang, Guomin <guomin.jiang@intel.com>
> > Sent: Thursday, April 23, 2020 9:34 AM
> > To: devel@edk2.groups.io; Jiang, Guomin <guomin.jiang@intel.com>; Ni,
> > Ray <ray.ni@intel.com>; Kinney, Michael D
> > <michael.d.kinney@intel.com>; ard.biesheuvel@linaro.org
> > Cc: lersek@redhat.com; macarl@microsoft.com
> > Subject: RE: [edk2-devel] [PATCH] CryptoPkg/FltUsedLib: Add FltUsedLib
> for float.
> >
> > The size comparation between with _fltused and without _fltused, use
> OvmfPkgX64 as build target
> > OvmfX64  EFI_FV_TAKEN_SIZE  |   Without _fltused       | With _fltused
> > DXEFV                                             |  0x46bbe0                     | 0x46bbc0
> > FVMAIN_COMPACT                     |  0x12c8a0                      | 0x12c7f0
> > PEIFV                                              |  0x4cae8                         | 0x4ca68
> > From the result, it will occupy less size rather than more size.
> >
> > Code Change as below and build command is ```build -p
> > OvmfPkg/OvmfPkgX64 -b DEBUG -t VS2019 -a X64 - DTPM_ENABLE=TRUE```
> >
> > diff --git a/CryptoPkg/Library/IntrinsicLib/MemoryIntrinsics.c
> > b/CryptoPkg/Library/IntrinsicLib/MemoryIntrinsics.c
> > index 94fe341bec..c4136916d0 100644
> > --- a/CryptoPkg/Library/IntrinsicLib/MemoryIntrinsics.c
> > +++ b/CryptoPkg/Library/IntrinsicLib/MemoryIntrinsics.c
> > @@ -2,7 +2,7 @@
> >    Intrinsic Memory Routines Wrapper Implementation for OpenSSL-based
> >    Cryptographic Library.
> >
> > -Copyright (c) 2010 - 2019, Intel Corporation. All rights
> > reserved.<BR>
> > +Copyright (c) 2010 - 2020, Intel Corporation. All rights
> > +reserved.<BR>
> >  SPDX-License-Identifier: BSD-2-Clause-Patent
> >
> >  **/
> > @@ -13,16 +13,6 @@ SPDX-License-Identifier: BSD-2-Clause-Patent
> >
> >  typedef UINTN  size_t;
> >
> > -#if defined(__GNUC__) || defined(__clang__)
> > -  #define GLOBAL_USED __attribute__((used)) -#else
> > -  #define GLOBAL_USED
> > -#endif
> > -
> > -/* OpenSSL will use floating point support, and C compiler produces the
> _fltused
> > -   symbol by default. Simply define this symbol here to satisfy the linker. */
> > -int  GLOBAL_USED _fltused = 1;
> > -
> >  /* Sets buffers to a specified character */  void * memset (void
> > *dest, int ch, size_t count)  { diff --git
> > a/MdePkg/Library/BaseLib/BaseLib.inf
> > b/MdePkg/Library/BaseLib/BaseLib.inf
> > index 3586beb0ab..bab31c07c7 100644
> > --- a/MdePkg/Library/BaseLib/BaseLib.inf
> > +++ b/MdePkg/Library/BaseLib/BaseLib.inf
> > @@ -1,7 +1,7 @@
> >  ## @file
> >  #  Base Library implementation.
> >  #
> > -#  Copyright (c) 2007 - 2019, Intel Corporation. All rights
> > reserved.<BR>
> > +#  Copyright (c) 2007 - 2020, Intel Corporation. All rights
> > +reserved.<BR>
> >  #  Portions copyright (c) 2008 - 2009, Apple Inc. All rights
> > reserved.<BR>  #  Portions copyright (c) 2011 - 2013, ARM Ltd. All
> > rights reserved.<BR>  # @@ -60,6 +60,7 @@
> >    String.c
> >    FilePaths.c
> >    BaseLibInternals.h
> > +  FltUsed.c               | MSFT
> >
> >  [Sources.Ia32]
> >    Ia32/WriteTr.nasm
> > diff --git a/MdePkg/Library/BaseLib/FltUsed.c
> > b/MdePkg/Library/BaseLib/FltUsed.c
> > new file mode 100644
> > index 0000000000..c065594266
> > --- /dev/null
> > +++ b/MdePkg/Library/BaseLib/FltUsed.c
> > @@ -0,0 +1,14 @@
> > +/** @file
> > +  Declare _fltused symbol for MSVC
> > +
> > +  MSVC need this symbol for float, andd it here to feed MSVC. it may
> > + remove  if MSVC not need it any more.
> > +
> > +  Copyright (c) 2020 - 2020, Intel Corporation. All rights
> > +reserved.<BR>
> > +  SPDX-License-Identifier: BSD-2-Clause-Patent **/
> > +
> > +//
> > +// Just for MSVC float
> > +//
> > +int _fltused = 0x1;
> >
> > > -----Original Message-----
> > > From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of
> > > Guomin Jiang
> > > Sent: Tuesday, April 14, 2020 3:01 PM
> > > To: devel@edk2.groups.io; Ni, Ray <ray.ni@intel.com>; Kinney,
> > > Michael D <michael.d.kinney@intel.com>; ard.biesheuvel@linaro.org
> > > Cc: lersek@redhat.com; macarl@microsoft.com
> > > Subject: Re: [edk2-devel] [PATCH] CryptoPkg/FltUsedLib: Add
> > > FltUsedLib for float.
> > >
> > > Summarize current status:
> > >
> > > Problem Statement:
> > > Openssl require _fltused to be defined as a constant anywhere
> > > floating point is used.
> > > It may use float out of edk2 tree and need _fltused, for example,
> > > Microsoft’s OnScreenKeyboard and UiToolKit.
> > >
> > > Current Proposal as below:
> > >
> > > Proposal 1: Add FltUsed.c into exist library
> > > Detail: Add FltUsed.c into EntryPointLib(PEIM, DXE).
> > > Recommend: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> > > Approve: Laszlo Ersek lersek@redhat.com
> > > Netual: Michael D Kinney <michael.d.kinney@intel.com>
> > > Benefit: Doesn’t need modify every .dsc description file.
> > > Defection: I test that it will fail because of /GL option, the error
> > > show fatal error LNK1237: during code generation, compiler
> > > introduced reference to symbol '_fltused' defined in module
> > > 'UefiApplicationEntryPoint.lib(FltUsed.obj)' compiled with /GL
> > >
> > > Proposal 2: Define NULL library
> > > Recommend: Michael D Kinney michael.d.kinney@intel.com
> > > Oppose: Sean sean.brogan=microsoft.com@groups.io , Ard Biesheuvel
> > > <ard.biesheuvel@linaro.org>
> > > Benefit: I test it and prove that it is executable.
> > > Defection: It is required that modify every description file.
> > > Defection: It need be very careful that it should only apply some
> > > module type(PEIM, DXE_DRIVER, UEFI_APPLICATION) rather than all.
> > > Defection: Build break up.
> > > Action Required: Need evaluate the affection on size.
> > > Consideration: Add PCD to control the feature
> > >
> > > Proposal 3: Define FltUsedLib
> > > Detail: Define FltUsedLib and add dependence
> > > Oppose: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> > > Benefit: Doesn’t need care that which module will use it, we will
> > > explicitly point it in component file.
> > > Defection: More complex, It is required that modify every
> > > description file and modify component meanwhile.
> > > Defection: Build breakup
> > >
> > > Thanks
> > > guomin
> > >
> > > > -----Original Message-----
> > > > From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of Ni,
> > > Ray
> > > > Sent: Tuesday, April 14, 2020 1:03 PM
> > > > To: devel@edk2.groups.io; Kinney, Michael D
> > > > <michael.d.kinney@intel.com>; ard.biesheuvel@linaro.org; Kinney,
> > > > Michael D <michael.d.kinney@intel.com>
> > > > Cc: lersek@redhat.com; macarl@microsoft.com
> > > > Subject: Re: [edk2-devel] [PATCH] CryptoPkg/FltUsedLib: Add
> > > > FltUsedLib for float.
> > > >
> > > > UEFI spec allows using float operation so asking module developers
> > > > avoid using float may not make sense. Even
> > > > UefiCpuPkg\Library\BaseUefiCpuLib\
> > > > provides routine to initialize float support for X86.
> > > >
> > > > Given ARM already uses NULL library class mechanism to supply the
> > > > compiler stub, I prefer X86 aligns to the ARM style.
> > > >
> > > > The only unsure thing is the size impact when a module is not
> > > > using float. We expect there is no size impact when a module is not
> using float.
> > > >
> > > > Thanks,
> > > > Ray
> > > >
> > > > > -----Original Message-----
> > > > > From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of
> > > > Michael
> > > > > D Kinney
> > > > > Sent: Thursday, April 2, 2020 12:38 AM
> > > > > To: devel@edk2.groups.io; ard.biesheuvel@linaro.org; Kinney,
> > > > > Michael D <michael.d.kinney@intel.com>
> > > > > Cc: lersek@redhat.com; macarl@microsoft.com
> > > > > Subject: Re: [edk2-devel] [PATCH] CryptoPkg/FltUsedLib: Add
> > > > > FltUsedLib for float.
> > > > >
> > > > > Hi Ard,
> > > > >
> > > > > I think adding a dependency in the module entry point libs is
> > > > > also good way to guarantee an intrinsic lib is available.  I
> > > > > agree that it can provide module type and arch specific
> > > > > mappings.  The NULL lib instance can do the same if the NULL
> > > > > instance is listed in the module/arch specific library classes
> > > > > sections. a few example section names.
> > > > >
> > > > > [LibraryClasses.ARM.PEIM, LibraryClasses.AARCH64.PEIM]
> > > > >
> > > > > [LibraryClasses.IA32.UEFI_DRIVER]
> > > > >
> > > > > [LibraryClasses.X64.DXE_DRIVER]
> > > > >
> > > > > So either way, the DSC files need to provide the library mapping.
> > > > > The only difference between these
> > > > > 2 approaches is that adding a dependency to the module entry
> > > > > point libs uses a formally defined library class name for the
> > > > > intrinsic functions vs.
> > > > > the un-named NULL library instance.  A formally defined library
> > > > > class name is better supported for things like unit tests from
> > > > > the UnitTestFrameworkPkg.
> > > > >
> > > > > The fltused issue would not go away of we removed the float
> > > > > usage for OpenSSL.  One or more other libs or the module could
> > > > > use float/double types and this issue would pop up again.  I do
> > > > > agree it would be cleaner if we could use OpenSSL without floating
> point.
> > > > >
> > > > > I think adding an intrinsic lib for IA32/X64 for VS20xx
> > > > > generation of intrinsic functions would address fltused and
> > > > > other common C implementation styles that generate intrinsic
> > > > > functions (e.g. 64-bit int math on 32-bit, structure variable
> > > > > assignments, and loops that fill a buffer with a const value) by
> > > > > VS20xx compilers.  An intrinsic lib for GCC would also help with
> > > > > these same common C implementation styles that generate intrinsic
> functions.
> > > > >
> > > > > Mike
> > > > >
> > > > > > -----Original Message-----
> > > > > > From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of
> > > Ard
> > > > > > Biesheuvel
> > > > > > Sent: Tuesday, March 31, 2020 11:43 PM
> > > > > > To: Kinney, Michael D <michael.d.kinney@intel.com>
> > > > > > Cc: devel@edk2.groups.io; lersek@redhat.com;
> > > > > > macarl@microsoft.com
> > > > > > Subject: Re: [edk2-devel] [PATCH] CryptoPkg/FltUsedLib:
> > > > > > Add FltUsedLib for float.
> > > > > >
> > > > > > On Tue, 31 Mar 2020 at 16:36, Kinney, Michael D
> > > > > > <michael.d.kinney@intel.com> wrote:
> > > > > > >
> > > > > > > ARM and AARCH64 have a compiler intrinsic lib that is
> > > > > > linked against all modules.
> > > > > > >
> > > > > > > [LibraryClasses.ARM, LibraryClasses.AARCH64]
> > > > > > >   #
> > > > > > >   # It is not possible to prevent ARM compiler calls to
> > > > > > generic intrinsic functions.
> > > > > > >   # This library provides the instrinsic functions
> > > > > > generated by a given compiler.
> > > > > > >   # [LibraryClasses.ARM] and NULL mean link this
> > > > > > library into all ARM images.
> > > > > > >   #
> > > > > > >
> > > > > > NULL|ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrins
> > > > > > icsLib.inf
> > > > > > >
> > > > > > > Can we use this same technique for IA32/X64 VS builds?
> > > > > > >
> > > > > >
> > > > > > In my opinion, having these intrinsics libraries added via
> > > > > > NULL library class resolution was a mistake to begin with.
> > > > > >
> > > > > > Every component that we build incorporates some kind of
> > > > > > startup code library that defines the _ModuleEntryPoint
> > > > > > symbol, and it would be much better to make those libraries
> > > > > > include an IntrinsicsLibrary dependency that can be satisfied
> > > > > > by arch specific versions that also encapsulate the toolchain
> > > > > > dependencies (such as this _fltused symbol, or memcpy/memset
> on ARM/GCC).
> > > > > >
> > > > > > On another note, it is still deeply disappointing that we need
> > > > > > to jump through all of these hoops because the *only* single
> > > > > > use of floating point in OpenSSL is the entropy estimate of an
> > > > > > RNG, which is in the range of 0..1023 to begin with [IIRC).
> > > > > > Remember that we also have a softfloat submodule for 32-bit
> > > > > > ARM, for the same stupid reason. If we could stop pulling in
> > > > > > that part of the code (or replace it with an
> > > > > > UINT16 when building for the UEFI target), the whole floating
> > > > > > point issue would mostly go away AFAICT.
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > 
> >
>