[llvm] r308319 - [TRE] Simplify canTRE() a bit using all_of(). NFCI.

Davide Italiano via llvm-commits llvm-commits at lists.llvm.org
Tue Jul 18 21:13:10 PDT 2017


On Tue, Jul 18, 2017 at 6:15 PM, Chris Lattner <sabre at nondot.org> wrote:
>
> On Jul 18, 2017, at 9:00 AM, Davide Italiano <davide at freebsd.org> wrote:
>
> On Tue, Jul 18, 2017 at 8:42 AM, Davide Italiano via llvm-commits
> <llvm-commits at lists.llvm.org> wrote:
>
> Author: davide
> Date: Tue Jul 18 08:42:59 2017
> New Revision: 308319
>
> URL: http://llvm.org/viewvc/llvm-project?rev=308319&view=rev
> Log:
> [TRE] Simplify canTRE() a bit using all_of(). NFCI.
>
> This has a ~11 years old FIXME, which may not be true today.
> We might consider removing this code altogether.
>
>
> Chris, any chance we can reconsider this?
> https://bugs.llvm.org/show_bug.cgi?id=962
>
>
> Hi Davide,
>
> What’s the specific question?
>
> -Chris

Oh, I replied on the PR, but let me elaborate here (apologies if I
wasn't clear).
I assume you don't have anymore the C/C++ code that caused the poor
codegen, and I wasn't able to trigger it from several programs I
built. At this point, I'm not entirely sure whether it's still
relevant special casing for dynamic allocas in TRE (although the
codegen is not great). That said, I thought about this a little more
and probably it's easier to keep the status quo as there's some risk
of regressing.

Thanks,

-- 
Davide

"There are no solved problems; there are only problems that are more
or less solved" -- Henri Poincare


More information about the llvm-commits mailing list