Re: [PATCH] OvmfPkg/Sec: Fix 64bit SEC build failure

[Laszlo Ersek <lersek@redhat.com>]

Hi Ray,

adding Ard, Jordan, Liming and Yonghong:

On 11/27/17 02:38, Ruiyu Ni wrote:
> Original code breaks a single assembly code to multiple lines.
> But build tool doesn't support such usage in Windows OS environment.
> 
> Changing the multiple lines to one line to resolve the build failure.
> 
> Contributed-under: TianoCore Contribution Agreement 1.1
> Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
> Cc: Laszlo Ersek <lersek@redhat.com>
> ---
>  OvmfPkg/Sec/X64/SecEntry.nasm | 6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
> 
> diff --git a/OvmfPkg/Sec/X64/SecEntry.nasm b/OvmfPkg/Sec/X64/SecEntry.nasm
> index 7c55032ac9..d76adcffd8 100644
> --- a/OvmfPkg/Sec/X64/SecEntry.nasm
> +++ b/OvmfPkg/Sec/X64/SecEntry.nasm
> @@ -45,10 +45,8 @@ ASM_PFX(_ModuleEntryPoint):
>      ; Fill the temporary RAM with the initial stack value.
>      ; The loop below will seed the heap as well, but that's harmless.
>      ;
> -    mov     rax, (FixedPcdGet32 (                        \
> -                    PcdInitValueInTempStack              \
> -                    ) << 32) |                           \
> -                 FixedPcdGet32 (PcdInitValueInTempStack)      ; qword to store
> +    mov     rax, (FixedPcdGet32 (PcdInitValueInTempStack) << 32) | FixedPcdGet32 (PcdInitValueInTempStack)
> +                                                              ; qword to store
>      mov     rdi, FixedPcdGet32 (PcdOvmfSecPeiTempRamBase)     ; base address,
>                                                                ;   relative to
>                                                                ;   ES
> 

I'd like to understand more about the build failure. As far as I understand, the NASM tool itself has no problem with the backslash line continuation; before I posted the patch, I checked the NASM manual for this feature.

http://www.nasm.us/doc/nasmdoc3.html#section-3.1

    NASM uses backslash (\) as the line continuation character; if a
    line ends with backslash, the next line is considered to be a part
    of the backslash-ended line.

Now, in "BaseTools/Conf/build_rule.template", I see that we have the following rule for NASM source files:

[Nasm-Assembly-Code-File.COMMON.COMMON]
    <InputFile>
        ?.nasm

    <ExtraDependency>
        $(MAKE_FILE)

    <OutputFile>
        $(OUTPUT_DIR)(+)${s_dir}(+)${s_base}.obj

    <Command>
        "$(PP)" $(PP_FLAGS) $(INC) ${src} > ${d_path}(+)${s_base}.i
        Trim --trim-long --source-code -o ${d_path}(+)${s_base}.iii ${d_path}(+)${s_base}.i
        "$(NASM)" -I${s_path}(+) $(NASM_FLAGS) -o $dst ${d_path}(+)${s_base}.iii

I guess it's actually the pre-processor (PP) that splices the broken-up lines together, as any C language pre-processor is supposed to do.

Why does that not work with the VS toolchains ("cl.exe")?

Is it perhaps a problem with the _PP_FLAGS?

https://msdn.microsoft.com/en-us/library/3xkfswhy.aspx
https://msdn.microsoft.com/en-us/library/032xwy55.aspx

"/E" seems right (pre-process). Is /TC the problem? ("C language source code").

If the problem is impossible to fix in BaseTools, I don't mind the patch, but I'd like to see more explanation in the commit message.

(Other assembly files use backslash line continuations too, they just haven't been exposed to VS toolchains yet, apparently.)

Thanks
Laszlo