Re: Yet another KPTI regression with 4.14.x series in a VM

[Thomas Gleixner <tglx@linutronix.de>]

On Sat, 13 Jan 2018, Peter Zijlstra wrote:

> On Fri, Jan 12, 2018 at 10:08:20PM -0800, Andy Lutomirski wrote:
> > Now this is quite a strange value to write to CR3.  The 0x800 part
> > means that we're using the "user" variant of the address space that
> > would have ASID=0 and the 0x1000 bit being set corresponds to the user
> > pgdir, but this is nonsense, since the kernel never uses PCID 0 for
> > user mode.  We always start at 1.  The only exception is if
> > X86_FEATURE_PCID is off.  But, if X86_FEATURE_PCID is off, then we
> > shouldn't be setting any PCID bits.
> 
> My bad, I was under the impression the lower 12 bits would be ignored
> without PCID :/

 2:0 Ignored
   3 PWT
   4 PCD
11:5 Ignored

So yes, it's mostly ignored at least in theory... I'm sure I stared at that
code and the SDM more than once and convinced myself that it's not an issue
to set bit 11 unconditionally.

But I should have stared at the AMD manual which says:

Reserved Bits. Reserved fields should be cleared to 0 by software when writing CR3.

Thanks,

	tglx