Those bits are reserved for fast math flags in all floating point operations,
though they only are relevant for binary ops. The setting is safe, but we could
have an assert if desired.
On Apr 6, 2014, at 3:58 AM, Blank, Guy <guy.blank at intel.com> wrote:
> Hi,
>
> I’ve noticed that the lang ref defines 5 operations (fadd, fsub, fmul,
fdiv, frem) that can use the fast-math flags.
> But when setting the flags in code, the only check is
isa<FPMathOperator>(instruction), which returns true for any instruction
that returns float.
> Should I really be able to set the flags for operations which were not
mentioned in the langref?
>
> Thanks,
> Guy
> ---------------------------------------------------------------------
> Intel Israel (74) Limited
>
> This e-mail and any attachments may contain confidential material for
> the sole use of the intended recipient(s). Any review or distribution
> by others is strictly prohibited. If you are not the intended
> recipient, please contact the sender and delete all copies.
>
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20140419/ba878ca4/attachment.html>