PATCH: fix clang to emit correct addrspacecast for CUDA

[Jingyue Wu]

No AFAIK. We can add the generic address space into LangAS::ID. The problem
is though how the IR optimizer knows which number means the generic address
space.


On Mon, Mar 24, 2014 at 1:22 PM, Justin Holewinski
<jholewinski at nvidia.com>wrote:

>  On 03/24/2014 03:22 PM, Matt Arsenault wrote:
>
> On 03/24/2014 11:46 AM, Justin Holewinski wrote:
>
>   Of course, this also assumes address space 0 is generic.  This is
> currently true for the in-tree targets and CUDA/OpenCL support in
> Clang, but I don't believe its a set rule anywhere.
>
>  R600 wants this optimization, except the "generic" address space isn't
> 0, and is instead some other currently undefined number.
>
>
> Sounds like good justification for making this a target-independent pass,
> then.
>
> Jingyue, is there anything in clang now that would prevent the generic
> address space from being a number other than 0?
>
>
> -Matt
>
>
>
>  ------------------------------
>  This email message is for the sole use of the intended recipient(s) and
> may contain confidential information.  Any unauthorized review, use,
> disclosure or distribution is prohibited.  If you are not the intended
> recipient, please contact the sender by reply email and destroy all copies
> of the original message.
>  ------------------------------
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20140324/1613ca60/attachment.html>