Ah, I see. This is my instruction:
%lsr.iv125.us.ph = phi i32 [ %lsr.iv.next126.us, %limit.exit281.i.us ], [
sub (i32 0, i32 ptrtoint ([8 x [2 x i32]]* @Maxlogmap_Decode_next_state to
i32)), %for.cond195.preheader.i.us ]
So there is a ptrtoint inside a sub inside a phi.
I'm just going to write a pass that promotes/demotes the embedded instr,
shouldn't take long.
Thanks,
On Thu, Sep 20, 2012 at 3:33 PM, Eli Friedman <eli.friedman at
gmail.com>wrote:
> On Thu, Sep 20, 2012 at 3:18 PM, Ryan Taylor <ryta1203 at gmail.com>
wrote:
> > Curious, is there a pass/mechanism that exists to promote an
instruction
> > that is a value of a PHI?
> >
> > For example, if one of the incoming values is a sub(x, y) from BB2,
does
> > there exist something to promote this instruction back to BB2 (such
that
> tmp
> > = sub(x,y)) and the PHI replaces the sub instr with tmp?
> >
> > Or is the above the natural state that is then optimized into using
sub,
> is
> > there a call that does this that I can turn off?
>
> It sounds like you're talking about a ConstantExpr. Those don't
> normally get lowered until CodeGen. You could expand it at the IR
> level, but there isn't any builtin function to do that IIRC.
>
> -Eli
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20120920/31e96b93/attachment.html>