[LLVMdev] tfloat support for mingw-w64

[Martell Malone]

ping?

+ Reid

On Mon, Jul 27, 2015 at 5:10 PM, Martell Malone <martellmalone at gmail.com>
wrote:

> Hi,
>
> I've been hacking around something missing in the assemble for the
> mingw-w64 targets
> the tfloat variable.
>
> I did some research into the llvm sources and did see x86_fp80 which seems
> to be the same thing.
> Can we support the .tfloat variable or the alternative ?
> Or is it under another name?
> I've tried using .x86_fp80 instead but to no avail. :/
>
> Here is how tfloat is being used in mingw-w64 and how I am hacking around
> it with a very dirty .double
>
>
> Date: Wed, 1 Jul 2015 02:53:22 +0100
> Subject: [PATCH 3/4] clang fixes and hacks
> diff --git a/mingw-w64-crt/math/log1pl.S b/mingw-w64-crt/math/log1pl.S
> --- a/mingw-w64-crt/math/log1pl.S
> +++ b/mingw-w64-crt/math/log1pl.S
> @@ -16,7 +16,11 @@
>         -1 + sqrt(2) / 2 <= x <= 1 - sqrt(2) / 2
>         0.29 is a safe value.
>       */
> +#ifdef __clang__
> +limit:    .double 0.29  // must enable .x86_fp80 for windows :)
> +#else    // double is very dirrrrty
>  limit:    .tfloat 0.29
> +#endif
>      /* Please note:     we use a double value here.  Since 1.0 has
>         an exact representation this does not effect the accuracy
>         but it helps to optimize the code.  */
>
> Many Thanks
> Martell
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20150729/8f85eb96/attachment.html>