Re: [GIT PULL] xen /proc/mtrr implementation

[Jeremy Fitzhardinge <jeremy@goop.org>]

Ingo Molnar wrote:
>> That's it.  I could add any number of bizarre convolutions to 
>> achieve the same effect, but given that there's an existing 
>> interface that is exactly designed for what we want to achieve, I 
>> have to admit it didn't occur to me to do anything else.
>>     
>
> Exactly what is 'bizarre' about using the API defined by the _CPU_ 
> already, without adding any ad-hoc hypecall? Catch the dom0 WRMSRs, 
> filter out the MTRR indices - that's it.
>   

Well, the x86 world can't seem to decide what the ABI is supposed to be, 
which is why we have mtrr_ops in the first place.  Doing emulation at 
the MSR level means that I'd need to decide which MTRR interface we're 
emulating today and do that.

Yes, I realize that almost everyone is using the same Intel-like 
interface these days, but it does mean there's a level of fragility that 
doesn't exist if we just implement mtrr_ops.

There's some secondary issues which arise.  For example, the mtrr 
trimming test is meaningless in dom0 (the e820 is fake, so it doesn't 
make sense to compare it with the mtrrs); we currently avoid that 
because the test only happens if the mtrr vendor is Intel.  We would 
need to disable that test some other way.

    J