Re: [PATCH v4 00/11] remove component-specific logic for AVX builds

[Bruce Richardson <bruce.richardson@intel.com>]

On Tue, Mar 25, 2025 at 08:46:35AM +0100, David Marchand wrote:
> Hello Bruce,
> 
> On Wed, Mar 19, 2025 at 7:09 PM Bruce Richardson
> <bruce.richardson@intel.com> wrote:
> >
> > On Wed, Mar 19, 2025 at 05:29:30PM +0000, Bruce Richardson wrote:
> > > A number of libs and drivers had special optimized AVX2 and AVX512 code
> > > paths for performance reasons, and these tended to have copy-pasted
> > > logic to build those files. Centralise that logic in the main
> > > drivers/ and lib/ meson.build files to avoid duplication.
> > >
> > > v4: rebase on latest main branch
> > >     minor fixes following feedback
> > >     limit use of -march=skylake-avx512 to when we don't already have a
> > >       -march flag supporting AVX512.
> > > v3: add patch for event/dlb2 AVX512 handling.
> > >     add common code for libraries as well as drivers.
> > > v2: add patch 4 to remove use of unnecessary CC_AVX2_SUPPORT flag
> > >
> > A related follow-up to this patchset. Checking with "godbolt.org", it
> > appears that both clang 3.6[1] and gcc 5[2] (the minimum called out compiler
> > versions in our docs[1]) support the set of AVX-512 compiler flags we use.
> > Therefore, it seems we can simplify our code further by removing the
> > "cc_has_avx512" variable.
> 
> What about https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90028 ?
> 

Yep, still needs to be handled, something I only realised after sending the
email.

> You'll need to send a new revision for this series in any case, since
> patch 9 broke the crc stuff in the net library.
> https://inbox.dpdk.org/dev/CAJFAV8w9wYPN+30Hv=batMvP=0M4momKZgmNDFixXbd-9u8t6w@mail.gmail.com/
> 
Yes, I saw that and just started looking at it last evening. Will hopefully
get a new revision out soon.

/Bruce