[PATCH] Add threaded variants of single-threaded malloc tests

[Collin Funk]

Mark Wielaard <mark@klomp.org> writes:

> On Wed, Nov 05, 2025 at 02:54:19PM -0500, DJ Delorie wrote:
>> Collin Funk <collin.funk1@gmail.com> writes:
>> > I would wait for more opinions from the original thread before putting
>> > things down in writing [1]. I haven't read the final binutils one yet.
>> 
>> That discussion was paused pending a discussion at Cauldron.  Did that
>> discussion happen?
>
> There was a lot of discussion around LLMs at the
> https://gcc.gnu.org/wiki/cauldron2025#Licensing_Birds_of_a_Feather
> Nick asked a lot of questions :)
>
> I like the binutils LLM policy, except for the "Generated-By:"
> part. It is IMHO super confusing to demand that for patches that are
> only acceptable because they are not legally significant.
>
> I do actually think it would be good to have a policy on
> machine/script/tool assisted/generated patches to make clear how a
> patch can be regenerated. But IMHO that should be a separate policy
> and discussion (and probably one that LLM generated patches cannot
> comply with because they cannot easily be regenerated).

The binutils policy looks fine to me. I would like to give Carlos a
chance to review it, though. As he mentioned here [1], we don't want to
disallow the use of assistive technology. I think the following line
makes that clear that it isn't the intention:

    Using LLMs to assist in writing code is fine providing that the LLM
    does not actually generate code. So for example using an LLM to
    provide text to speech services or to search for published
    information is OK.

But maybe another set of eyes will catch issues with that.

I agree with you on the "Generated-By:" part. I don't see the benefit,
and I rather avoid those extra lines in the 'git log'.

Thanks,
Collin

[1] https://inbox.sourceware.org/libc-alpha/146e5c2e-64e9-4531-bbec-72cb3b323df6@redhat.com/