|   | 
      | 
  
  
      | 
      | 
  
 
     | 
    | 
  
  
     | 
    | 
  
  
    |   | 
      | 
  
  
    | 
         
xen-devel
RE: [Xen-devel] [PATCH] [HVM] Rename file hvm_info_table.h	toplatform.h
 
>
>On 27 Jul 2006, at 11:25, Li, Xin B wrote:
>
>> Rename file hvm_info_table.h to platform.h.
>> We are adding more HVM platform definitions that will be used in 
>> control
>> panel, device model and hypervisor, so a new file platform.h will be
>> created to hold such definitions.  Actually, definitions in
>> hvm_info_table.h could be moved to this new file.
>>
>> Signed-off-by: Xin Li <xin.b.li@xxxxxxxxx>
>
>What kinds of things are going to be added? 'platform.h' is vague 
>enough it could end up a dumping ground for all kinds of crud.
>
To support HVM guest with RAM above 4G physical address space, we need
define a constant HVM_RAM_LIMIT_BELOW_4G, and physical address space
from HVM_RAM_LIMIT_BELOW_4G to 4G is reserved for PCI device MMIO use.
So if HVM guest has more than HVM_RAM_LIMIT_BELOW_4G RAM, RAM beyond
HVM_RAM_LIMIT_BELOW_4G should go to physical address space above 4G.  So
p2m table and e820 table need  adjust Accordingly.
The constant HVM_RAM_LIMIT_BELOW_4G will be used in control panel,
device model and hypervisor, and I need a header file to hold the
definition.  It's hard for me to find a good English name for the header
file.  Another concern is, in the future, we may have more such
definitions.
BTW, Qemu-dm allocates PCI devices MMIO from 0xf0000000.
-Xin
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
 
 |   
 
| <Prev in Thread] | 
Current Thread | 
[Next in Thread> |  
- RE: [Xen-devel] [PATCH] [HVM] Rename file hvm_info_table.h	toplatform.h,
Li, Xin B <=
  
 |  
  
 | 
    | 
  
  
    |   | 
    |