On Mon, 2011-09-26 at 15:38 -0400, Lennart Sorensen wrote: > On Mon, Sep 26, 2011 at 03:34:36PM -0400, Lennart Sorensen wrote: > > Ehm, how can PSE introduced by the Pentium III be pre-PAE (which was > > introduced by the Pentium Pro)? > > > > PSE may be simpler but it's still newer than PAE. > > Well actually PSE (support for larger than 4k pages) predates PAE, but > PSE-36 (support for larger than 4k pages with 36bit physical address > range support) comes after PAE. My mistake, I thought PSE-36 also pre-dated PAE but obviously I was wrong. One thing which is interesting is that "pse36" normally appears separate from "pse" in /proc/cpuinfo but doesn't appear here, just like "pae" is missing I don't see anywhere in the kernel which masks off X86_FEATURE_PAE or X86_FEATURE_PSE36 though. I guess running x86info would give the real underlying features untouched by the kernel if Bernhard wants to confirm. Ian. > > PSE came in the pentium, PAE in the Pentium Pro, and PSE-36 in the > Pentium III. > > -- > Len Sorensen > > > -- Ian Campbell Interchangeable parts won't.
Attachment:
signature.asc
Description: This is a digitally signed message part