[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <E828A2FA-A5DA-472E-9C22-C04AF91D13A0@oracle.com>
Date: Wed, 8 Oct 2025 14:09:51 +0000
From: Qing Zhao <qing.zhao@...cle.com>
To: Kees Cook <kees@...nel.org>, "claziss@...il.com" <claziss@...il.com>
CC: Claudiu Zissulescu <claziss@...il.com>,
"gcc-patches@....gnu.org"
<gcc-patches@....gnu.org>,
"linux-hardening@...r.kernel.org"
<linux-hardening@...r.kernel.org>
Subject: Re: [PATCH] arc: Add const attribute support for mathematical ARC
builtins
> On Oct 7, 2025, at 23:29, Kees Cook <kees@...nel.org> wrote:
>
> On Tue, Oct 07, 2025 at 09:02:21PM +0000, Qing Zhao wrote:
>> Hi, Kees,
>>
>> I took a look at the patch today.
>> Overall, I think the patch is good and should work well.
>
> Thanks!
>
>> My major questions are:
>>
>> 1. Are the five functions in the current list the only functions in __builtin_arc_*** that
>> have const attribute?
>
> They are the only ones I can safely assert are const attribute without
> doing additional work. I'm sure there are more, but I didn't feel
> comfortable marking others.
>
>> 2. In addition to “const” attribute, are there any other important attributes should be added for
>> the arc builtins to enable more optimizations?
>
> Again, I'm not sure. But my goal here is to make it possible to add
> these going forward if people find stuff to add. Before this patch it
> wasn't possible at all -- the infrastructure to do so was missing, so
> that's what I wanted to add: the infrastructure and the first user. :)
Yeah, understood. And sounds reasonable to me.
Does Claudiu have more concern and comments on this patch?
thanks.
Qing
>
> --
> Kees Cook
Powered by blists - more mailing lists