instruction 'align' in asm blocks works incorrectly with some cases of parameters clean revision

[m_zuckerman]

Summary

This is a clean revision:
The old revision is http://reviews.llvm.org/D13869#inline-115174
This align directive doesn't work on the -fasm-blocks.

ALIGN [[number]]
Aligns the next variable or instruction on a byte that is a multiple of number.

Microsoft support only power of 2 align like gcc and icc .
The problem with the code was in the AsmParser.cpp
Inside the AsmParser::parseDirectiveMSAlign function we trying to lowering number in to log of two. This is not correct and not consistent with Microsoft compilers.
**CASE 1: Microsoft compatibility
align should work only with address of multiply of two.

int f(void)
 {
__asm { align 2 }
__asm { align 4 }
__asm { align 8 }
__asm { align 16 }
__asm { align 32 }
return 0;


}

int main()
 {
return f();
}

CASE 2: Darwin compatibility
Incorrect lowering inline asm with -fasm-blocks for Darwin.

int f(void)
{
    asm { align 2 }
    asm { align 4 }
 
    return 0;
}

Output

## InlineAsm Start
.align 1, 0x90
## InlineAsm End
mov dword ptr [esp], eax

## InlineAsm Start
.align 2, 0x90
## InlineAsm End

It's should be

## InlineAsm Start
.align 2, 0x90
## InlineAsm End
mov dword ptr [esp], eax

## InlineAsm Start
.align 4, 0x90
## InlineAsm End

For not log2 number like align 3 error retrieve.

[m_zuckerman]

lib/MC/MCParser/AsmParser.cpp

4509 ↗

(On Diff #38546)

Add check for only In bytes mode for the number is actually log 2 of the number.

4513 ↗

(On Diff #38546)

Only need to rewrite the align N to .align N
In all OS.
Only Microsoft was lowering the align N to .align log2(N) in the past.
Now all OS lowering the align N to .align N

4679 ↗

(On Diff #38546)

Not need additional skip because no new value are pass.

4715–4716 ↗

(On Diff #38546)

There is no new val.

4718–4720 ↗

(On Diff #38546)

This was relevant when Microsoft lower the align N to .align log2N

test/CodeGen/ms-inline-asm.c

294–299 ↗

(On Diff #38546)

Incorrect lowering of the align N for Darwin The running mode is
// RUN: %clang_cc1 %s -triple i386-apple-darwin10 -fasm-blocks -emit-llvm -o - | FileCheck %s

[rnk]

test/CodeGen/ms-inline-asm.c

294–299 ↗

(On Diff #38546)

I don't think this is incorrect. This appears to be our fundamental disagreement. Please review my comments from earlier about how this should work.

llvm/trunk/lib/MC/MCParser/AsmParser.cpp

case AOK_SizeDirective:
       case 128: OS << "xmmword ptr "; break;
       case 256: OS << "ymmword ptr "; break;
       }
       break;
     case AOK_Emit:
       OS << ".byte";
       break;
     case AOK_Align: {
-      unsigned Val = AR.Val;
-      OS << ".align " << Val;
+      // MS alignment directives are measured in bytes. If the native assembler
+      // measures alignment in bytes, we can pass it straight through.
+      OS << ".align";
+      if (getContext().getAsmInfo()->getAlignmentIsInBytes())
+        break;
 
-      // Skip the original immediate.
+      // Alignment is in log2 form, so print that instead and skip the original
+      // immediate.
+      unsigned Val = AR.Val;
+      OS << ' ' << Val;
       assert(Val < 10 && "Expected alignment less then 2^10.");
       AdditionalSkip = (Val < 4) ? 2 : Val < 7 ? 3 : 4;
       break;
     }
     case AOK_DotOperator:
       // Insert the dot if the user omitted it.
       OS.flush();
       if (AsmStringIR.back() != '.')