On 10.03.2026 16:29, Andrew Cooper wrote: > On 10/03/2026 10:26 am, Jan Beulich wrote: >> On 04.03.2026 20:53, Kevin Lampis wrote: >>> struct cpuinfo_x86 >>> .x86 => .family >>> .x86_vendor => .vendor >>> .x86_model => .model >>> .x86_mask => .stepping >>> >>> No functional change. >>> >>> This work is part of making Xen safe for Intel family 18/19. >>> >>> Signed-off-by: Kevin Lampis <[email protected]> >>> --- >>> I couldn't find any information about these Intel family 6 processors: >>> model 0x5d: /* SoFIA 3G Granite/ES2.1 */ >>> model 0x65: /* SoFIA LTE AOSP */ >>> model 0x6e: /* Cougar Mountain */ >>> Should I add them to intel-family.h? >> We should keep that in sync with Linux, I think, so "no" unless you mean to >> first have respective entries added there. >> >> 0x5d is listed in the SDM, so I'd recommend keeping the entries (using IFM() >> directly for the time being). >> >> 0x65 and 0x6e aren't listed in the SDM, so may never really have hit the >> public. Respective cases can perhaps be dropped, but such dropping would >> again likely better be done in a separate change. > > These are the 5G basestations I mentioned during the x86 meeting. Linux > has declined to take take these models into intel-family.h because > they're not general purpose. > > I suggest we drop them, including 0x5d. I can do a separate patch.
For 0x5d it's not quite clear what the justification would be. Looking at the SDM, those X3-C3000 don't look any more or less general purpose as C2000 and C1000. I guess I'll learn once you've sent the patch. Jan
