| 85 |
|
|
| 86 |
|
/* |
| 87 |
|
* Helpers to share 32-bit addressable data with MacOS |
| 88 |
+ |
* |
| 89 |
+ |
* There are two distinct allocatable regions: |
| 90 |
+ |
* |
| 91 |
+ |
* - The Data region is used to share data between MacOS and |
| 92 |
+ |
* SheepShaver. This is stack-like allocation since it is |
| 93 |
+ |
* meant to only hold temporary data which dies at the end |
| 94 |
+ |
* of the current function scope. |
| 95 |
+ |
* |
| 96 |
+ |
* - The Procedure region is used to hold permanent M68K or |
| 97 |
+ |
* PowerPC code to assist native routine implementations. |
| 98 |
+ |
* |
| 99 |
+ |
* - The Procedure region grows up whereas the Data region |
| 100 |
+ |
* grows down. They may intersect into the ZeroPage, which |
| 101 |
+ |
* is a read-only page with all bits set to zero. In practise, |
| 102 |
+ |
* the intersection is unlikely since the Procedure region is |
| 103 |
+ |
* static and the Data region is meant to be small (< 256 KB). |
| 104 |
|
*/ |
| 105 |
|
|
| 106 |
|
class SheepMem { |
