[llvm-commits] [llvm] r58742 - /llvm/trunk/lib/Transforms/Utils/InlineCost.cpp

[Daniel Dunbar]

I don't understand the motivation for this change (test case?), although it
seems fine.

Do we still need to compute the NeverInline set in InlineAlways.cpp though?

 - Daniel

On Tue, Nov 4, 2008 at 5:37 PM, Devang Patel <dpatel at apple.com> wrote:

> Author: dpatel
> Date: Tue Nov  4 19:37:05 2008
> New Revision: 58742
>
> URL: http://llvm.org/viewvc/llvm-project?rev=58742&view=rev
> Log:
> Check Attribute::NoInline.
>
> Modified:
>    llvm/trunk/lib/Transforms/Utils/InlineCost.cpp
>
> Modified: llvm/trunk/lib/Transforms/Utils/InlineCost.cpp
> URL:
> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Utils/InlineCost.cpp?rev=58742&r1=58741&r2=58742&view=diff
>
>
> ==============================================================================
> --- llvm/trunk/lib/Transforms/Utils/InlineCost.cpp (original)
> +++ llvm/trunk/lib/Transforms/Utils/InlineCost.cpp Tue Nov  4 19:37:05 2008
> @@ -182,11 +182,9 @@
>       // FIXME: We allow link-once linkage since in practice all versions
> of
>       // the function have the same body (C++ ODR) - but the LLVM
> definition
>       // of LinkOnceLinkage doesn't require this.
> -      (Callee->mayBeOverridden() && !Callee->hasLinkOnceLinkage()
> -       ) ||
> -
> +      (Callee->mayBeOverridden() && !Callee->hasLinkOnceLinkage()) ||
>       // Don't inline functions marked noinline.
> -      NeverInline.count(Callee))
> +      Callee->hasFnAttr(Attribute::NoInline) || NeverInline.count(Callee))
>     return llvm::InlineCost::getNever();
>
>   // InlineCost - This value measures how good of an inline candidate this
> call
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20081107/daf7dbc5/attachment.html>