[eros-arch] Problem with the semantics of address translation and
clandau at macslab.com
Fri Sep 2 18:37:08 EDT 2005
At 3:05 PM -0700 9/2/05, Norman Hardy wrote:
>This architecture is unfortunate in that it is an awkward machine dependency.
>It was easy to implement.
>Perhaps the architecture should merely say for some fixed n that
>"there is a limit of at least n levels on each portion".
One way to specify this that is more architecture-independent is to
say that a path length of up to n is permitted on each portion of the
tree going from one level (SSC/BLSS) to the next lower level,
regardless of whether a mapping table occurs at that level. The
implementation might only check at levels with mapping tables. For
example, from SSC 6 to SSC 4 it would check that the path length is
less than or equal to 2*n.
More information about the eros-arch