On Wed, Jun 12, 2019 at 07:51:26AM -0700, Dave Hansen wrote: > On 6/12/19 2:52 AM, Andy Shevchenko wrote: > >> #define INTEL_FAM6_ICELAKE_MOBILE 0x7E > >> +#define INTEL_FAM6_ICELAKE_NNPI 0x9D > > What "I" stands for? > > > > For me sounds like it's redundant here or something like NNP_DLI would be > > better (because somewhere we have _NP as for Network Processor). > > Let's not bikeshed this too much. These things aren't used that widely > and mostly they're just used for figuring out the processor generation. > It's exceedingly rare to have something like: > > > if (model == INTEL_FAM6_ICELAKE_MOBILE) > foo(); > else if (model == INTEL_FAM6_ICELAKE_NNPI) > bar(); > > where what you suggest would matter. > > Preserving the ability to google "ice lake nnpi" is pretty important, so > preserving the Intel name makes a lot of sense here when possible. What I'm talking is a consistency among suffixes. If there is a real abbreviation (NNPI) which anybody can google, I'm totally for it to be used. > Do we *HAVE* an Ice Lake network processor? Not Ice Lake, something else. -- With Best Regards, Andy Shevchenko
