residency properties and covert channels

[Neal H. Walfield]

At Sat, 05 Nov 2005 12:49:46 -0500,
Jonathan S. Shapiro wrote:
>   1. Self-paging introduces a known, surprisingly high-bandwidth
>      communication channel.
> 
>      Actually, this is not correctly stated. There is no problem with
>      paging against your own pages and deciding which of your pages to
>      unbind when you do. The problem lies in any system where the system
>      asks you to alter your residency properties -- this is the source
>      of the communication channel.

I'm having difficulty understanding the implication of this
communication channel.  It seems to me that using the L1 or L2 cache
would be a much higher bandwidth communication channel.  What I'm
thinking of is that two confined tasks, A and B, want to communicate.
So, A floods a cache color and B samples the different colors to
determine which color is least responsive.

Basically, I don't understand why this particular channel is so bad.

Thanks,
Neal