PAE is_guest_l2_slot()

PAE is_guest_l2_slot()

[Jan Beulich]

Could anyone help me understand what the meaning of the hardcoded 3 in that
macro is? With PGT_va_mask being an 11-bit mask I can't seem to make any sense
of this.
And while at this, for 64-bits PGT_va_mask is defined as a 28-bit mask, but the
comment there as well as my calculations say this should be only 27 bits.

Thanks, Jan

Re: PAE is_guest_l2_slot()

[Gerd Hoffmann]

Jan Beulich wrote:
> Could anyone help me understand what the meaning of the hardcoded 3 in that
> macro is?

xen address space window is in l2 table #3 (of 4) ...

cheers,

  Gerd

-- 
Gerd Hoffmann <kraxel@suse.de>

Re: PAE is_guest_l2_slot()

[Jan Beulich]

>>> Gerd Hoffmann <kraxel@suse.de> 12.09.06 16:25 >>>
>Jan Beulich wrote:
>> Could anyone help me understand what the meaning of the hardcoded 3 in that
>> macro is?
>
>xen address space window is in l2 table #3 (of 4) ...

That's why I pointed out that the mask is 11 bits (not 2) - what you say would mean
that the constant should be 0x600 (the top 11 bits of such a va), and comparison
should be >= rather than ==.

Jan

Re: PAE is_guest_l2_slot()

[Gerd Hoffmann]

Jan Beulich wrote:
>>>> Gerd Hoffmann <kraxel@suse.de> 12.09.06 16:25 >>>
>> Jan Beulich wrote:
>>> Could anyone help me understand what the meaning of the hardcoded 3 in that
>>> macro is?
>> xen address space window is in l2 table #3 (of 4) ...
> 
> That's why I pointed out that the mask is 11 bits (not 2) - what you say would mean
> that the constant should be 0x600 (the top 11 bits of such a va), and comparison
> should be >= rather than ==.

IIRC the full 11 bits are needed for L1 tables only (2 indexing into L3
and 9 indexing into L2).  Has been quite some time I looked into this
code ...

cheers,

  Gerd


-- 
Gerd Hoffmann <kraxel@suse.de>