Re: [patch] UefiCpuPkg/PiSmmCpuDxeSmm: Add "extern" keyword for "gPatchxxx"

[Laszlo Ersek <lersek@redhat.com>]

Hello Dandan,

On 04/12/18 10:50, Dandan Bi wrote:
> Background description:
> In SmmProfileInternal.h, ECC check tool report an issue at line 103.
> Detailed ECC Error info:Variable definition appears in header file.
> Include files should contain only public or only private data and
> cannot contain code or define data variables
>
> ECC report similar issues in PiSmmCpuDxeSmm.h.
>
> Then we review all the new introduced "gPatchxxx", since they have
> been defined in the nasm file, we can add "extern" keyword for them
> in the C source or header files.
>
> Cc: Eric Dong <eric.dong@intel.com>
> Cc: Laszlo Ersek <lersek@redhat.com>
> Contributed-under: TianoCore Contribution Agreement 1.1
> Signed-off-by: Dandan Bi <dandan.bi@intel.com>
> ---
>  UefiCpuPkg/PiSmmCpuDxeSmm/PiSmmCpuDxeSmm.h     | 8 ++++----
>  UefiCpuPkg/PiSmmCpuDxeSmm/SmmProfileInternal.h | 2 +-
>  UefiCpuPkg/PiSmmCpuDxeSmm/SmramSaveState.c     | 6 +++---
>  UefiCpuPkg/PiSmmCpuDxeSmm/X64/Semaphore.c      | 4 ++--
>  4 files changed, 10 insertions(+), 10 deletions(-)

This is a bug (a false positive) in the ECC tool. The following
declaration:

> X86_ASSEMBLY_PATCH_LABEL            gPatchSmmCr0;

does not declare an *object* (a variable). Instead, it declares a
*function* (and not a pointer to a function!), because (from
"MdePkg/Include/Library/BaseLib.h"):

> ///
> /// Type definition for representing labels in NASM source code that allow for
> /// the patching of immediate operands of IA32 and X64 instructions.
> ///
> /// While the type is technically defined as a function type (note: not a
> /// pointer-to-function type), such labels in NASM source code never stand for
> /// actual functions, and identifiers declared with this function type should
> /// never be called. This is also why the EFIAPI calling convention specifier
> /// is missing from the typedef, and why the typedef does not follow the usual
> /// edk2 coding style for function (or pointer-to-function) typedefs. The VOID
> /// return type and the VOID argument list are merely artifacts.
> ///
> typedef VOID (X86_ASSEMBLY_PATCH_LABEL) (VOID);

That is, when you see

> X86_ASSEMBLY_PATCH_LABEL            gPatchSmmCr0;

That is identical to the following function declaration:

> VOID gPatchSmmCr0 (VOID);

Now, the ISO C99 standard says:

> 6.2.2 Linkages of identifiers
>
> [...]
>
>   5 If the declaration of an identifier for a function has no
>     storage-class specifier, its linkage is determined exactly as if
>     it were declared with the storage-class specifier /extern/. [...]

Thus, the report from ECC is a false positive.

I don't mind the patch (the changes don't make any difference at the
C-language level, see the spec above); however, the commit message
should be 100% clear that the patch works around a limitation with the
ECC tool.

Can you please submit v2 with an updated commit message?

Thanks!
Laszlo