On Thu, Aug 22, 2002 at 11:58:37AM -0700, Jason R Thorpe wrote:
> Now, Chuq *claims* that it's also okay to have an aliasing writable
> unmanaged mapping if you don't actually write to it.  However, I am
> skeptical of this.  The NFS read-ahead code seems to do this (I've
> been trying to track down a data corruption bug on XScale that happens
> only when using the read/write-allocate cache line policy in conjuction
> with NFS).


well, it could turn out that there's a problem there, but I don't know
what it would be.

the idea is that it's the cache state that can cause problems, not
the TLB state.  as long as all of the duplicate cache lines have the
same data, everything should be happy.  the ARM pmap already allows
this for multiple read-only mappings, so I figure it should be ok
for writable mappings too as long as you only read from them.
or in the read-ahead case, it doesn't actually dereference the
mappings to already-initialized pages at all, it just creates them
and removes them later.  yeayea, it's kinda silly, but it shouldn't
cause any problems.

-Chuck