[PATCH] D70157: Align branches within 32-Byte boundary

[Fedor Sergeev via Phabricator via llvm-commits]

fedor.sergeev added a comment.

In D70157#1774561 <https://reviews.llvm.org/D70157#1774561>, @skan wrote:

> I could not reproduce the phenomenon that N-byte nop becomes (N+M) bytes with your example. So according to my understanding, I slightly modified your case. (If my understand is wrong, I hope you can point it out :-). )
>
>       .text
>       nop
>   .Ltmp0:
>       .p2align 3, 0x90
>       .rept 16
>       nop
>       .endr
>   .Ltmp3:
>       movl  %eax, -4(%rsp)


In our case it was

  andl $1, %eax

but it does not matter that much.

>   .rept 2
>   nop
>   .endr
>   jmp .Ltmp0
> 
>   The instruction `jmp .Ltmp0` starts at byte 0x1e and ends at byte 0x20.

Again, in our particular case start of the sequence was at xxx8, so 8 + 16(our sequence) + 3(andl) + 5(jmp) == 32.

> If we align the jump with preifx, two prefixes will be added to the `.rept2 16 nop .endr`. After prefixes are added, the 16-byte nop becomes 18-byte nop, then the label '.Ltmp3' is not 8-byte aligned any more.

Yes, thats what happened.

> I doubt whether the assumption that '.Ltmp3' is 8-byte aligned is right, since the alignment is not explicitly required.

The point is that we have explicit requirement at the start and we have a lowering into 16-byte sequence that we need to be preserved exactly as it is.
Essentially what we need is  a "protection" for this sequence from any changes by machinery that generates the binary code.
How can we protect a particular byte sequence from being changed by this branch aligner?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D70157/new/

https://reviews.llvm.org/D70157