> -----Original Message-----
> From: David Marchand <david.march...@redhat.com>
> Sent: Sunday, November 1, 2020 12:01 PM
> To: McDaniel, Timothy <timothy.mcdan...@intel.com>
> Cc: dev <dev@dpdk.org>; Carrillo, Erik G <erik.g.carri...@intel.com>; Eads,
> Gage <gage.e...@intel.com>; Van Haaren, Harry
> <harry.van.haa...@intel.com>; Jerin Jacob Kollanukkaran
> <jer...@marvell.com>; Thomas Monjalon <tho...@monjalon.net>
> Subject: Re: [dpdk-dev] [PATCH v14 05/23] event/dlb: add inline functions
>
> > > > +
> > > > +static inline void
> > > > +dlb_movdir64b(void *dest, void *src)
> > > > +{
> > > > + asm volatile(".byte 0x66, 0x0f, 0x38, 0xf8, 0x02"
> > > > + :
> > > > + : "a" (dest), "d" (src));
> > > > +}
> > >
> > > NO!
> > > We introduced stuff in EAL for this, please double check.
> > I also do not see a replacement for the new MOVDIR64B instruction in dpdk-
> next-eventdev or dpdk main.
>
> Ok, what got introduced in EAL is for MOVDIRI.
>
>
> So here we go with a MOVDIR64B...
> Google tells me:
> Availability of the MOVDIR64B instruction is indicated by the presence
> of the CPUID feature flag MOVDIR64B (bit 28 of the ECX register in
> leaf 07H, see “CPUID—CPU Identification” in the Intel® 64 and IA-32
> Architectures Software Developer’s Manual, Volume 2A).
>
> I understand that calling this code must be under a check for
> RTE_CPUFLAG_MOVDIR64B.
> Which I can't find in this patchset.
> What did I miss this time?
>
>
> --
> David Marchand
Fair enough question. We currently do not check for availability of MOVDIR64B
since
every Intel part that includes DLB or DLB2 is guaranteed to have the MOVDIR64B
instruction.
