Re: [PATCH 7/7] KVM: MMU: cache guest page number to guest frame number

[Avi Kivity <avi@redhat.com>]

On 02/23/2011 03:38 AM, Xiao Guangrong wrote:
> On 02/22/2011 10:32 PM, Avi Kivity wrote:
> >  On 02/22/2011 10:16 AM, Xiao Guangrong wrote:
> >>  Cache guest page number to guest frame number to avoid walk guest page table
> >>  frequently, the 'vtlb' idea is from Xen.
> >>
> >>  Note:
> >>  we can't use vtlb in ept guests since the guest tlb invalid operation is not
> >>  intercept(reload CR3, invlpg), also can't used in L2 nnpt guest for the same
> >>  reason, but we can used it to cache L1's npt page table.
> >>
> >
> >  I'm not so hot about introducing a new mechanism strictly for older hosts... EPT exists in three generations of Intel processors now (Sandy Bridge, Westmere, and Nehalem), and NPT is significantly older.
> >
>
> Um...so, do we should stop the new features for softmmu, only bug fix is welcome? :-)

No.  There is always a tradeoff between features and complexity.  What 
I'm saying is that I want to shift the tradeoff, for older processors, 
towards reducing complexity.  An improvement that is very simple, or 
gives very large gains, will be accepted.  A complex improvement that 
gives small gains may be rejected (but if it's for newer processors, it 
may be accepted).  It's a way for the maintainers to manage the ever 
growing complexity.

-- 
error compiling committee.c: too many arguments to function