Re: [PATCH] x86_64: Implement AVX2 version of strlcpy/wcslcpy function

[Noah Goldstein <goldstein.w.n@gmail.com>]

On Fri, Jun 30, 2023 at 6:22 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
>
> On Fri, Jun 30, 2023 at 5:21 PM Sunil Pandey <skpgkp2@gmail.com> wrote:
> >
> >
> >
> > On Fri, Jun 30, 2023 at 2:27 PM Paul Eggert <eggert@cs.ucla.edu> wrote:
> >>
> >> On 2023-06-30 14:04, Noah Goldstein via Libc-alpha wrote:
> >> > Think we should at the very least wait for the generic strlcpy codes
> >> > to land first.
> >>
> >> Let's not optimize these functions at all, unless there's good and
> >> measured reason to do so. In practice I expected they're called with
> >> small sizes for which optimization is a net minus as it consumes
> >> valuable maintenance time with no real benefit.
> >
> >
> > Hi Paul,
> >
> > Attached is strcpy/wcslcpy microbenchmark data based on Noah strlcpy/wcslcpy microbenchmark patch.
> >
> I don't think the concern is that we can beat the generic impl (which hasn't
> even landed yet AFAICT), it whether doing so makes sense given the
> usage/goal of the functions.
>

That being said, I'm generally in favor of adding optimized versions since
we happen to be a position where at least several developers find it worth
their time to maintain, but not before the generic versions have landed.
> > https://sourceware.org/pipermail/libc-alpha/2023-April/147557.html
> >
> > Thanks,
> > Sunil