Re: [PATCH] net: stop using mmx intrinsics

[Tyler Retzlaff <roretzla@linux.microsoft.com>]

On Thu, Mar 21, 2024 at 07:01:17PM +0100, Thomas Monjalon wrote:
> 21/03/2024 18:27, Tyler Retzlaff:
> > On Thu, Mar 21, 2024 at 06:09:01PM +0100, Thomas Monjalon wrote:
> > > 20/03/2024 22:12, Tyler Retzlaff:
> > > > +#ifdef RTE_TOOLCHAIN_MSVC
> > > > +#include <intrin.h>
> > > > +#else
> > > >  #include <x86intrin.h>
> > > > +#endif
> > > 
> > > It is not the same include in MSVC?
> > 
> > unfortunately intrin.h is vestigial in the monolithic approach. to use
> > any intrinsic you're supposed to include only the one and only true
> > header instead of vendor/arch feature specific headers.
> > 
> > > Is it something we want to wrap in a DPDK header file?
> > 
> > do you mean create a monolithic rte_intrinsic.h header that is
> > essentially
> > 
> > #ifdef MSVC
> > #include <intrin.h>
> > #else
> > #include <x86intrin.h>
> > #include <immintrin.h>
> > #include <nmmintrin.h>
> > ...
> > #endif
> > 
> > i assumed that doing something like this might be unpopular due to the
> > unnecessary namespace pollution.
> 
> We already have such a file.
> It is rte_vect.h.
> I suppose we should just make sure it is included consistently
> instead of x86intrin.h or immintrin.h
> 
> This command will show where changes are required:
> 	git grep intrin.h

there were some corner cases i can't recall, but since you identified
rte_vect.h is the preferred header let me do some experiments to see
what i can learn.  i'll either submit a series addressing it
specifically or come back with details.

thanks!

> 
>