These are typically I/O device register accesses in full-system mode.
I don't believe you'll see any of these in syscall emulation.

Steve

On Fri, May 28, 2010 at 12:39 AM, Malek Musleh <[email protected]> wrote:
> Hello,
>
> I was wondering what is the significance / implementation of the
> boolean function unCacheable() defined in the request.hh structure. I
> see that it is used the MSHR implementation as well as in some other
> instances, and i see that the corresponding comment in the request.hh
> file says that it indicates whether or not the request to the
> specified address is uncacheable, no additional comments found.
>
> How is / would an address be uncacheable? Does it correlate to an
> address possibly in memory or just a cache? I don't see why an address
> would not be cacheable in a cache, since replacement policy dictates
> the storage of addresses/data there.
>
> I guess I could see this being useful in replicating/indicating
> addresses volatile variables (variables in which local copies are not
> made to each private cache of a sharing core in order to minimize
> coherence traffic/issues.
>
> Thanks.
>
> Malek
> _______________________________________________
> m5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
>
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users

Reply via email to