>>>>> "TM" == Trond Myklebust <trond.myklebust@fys.uio.no> writes:

EB> And this still leaves it as an open question why keep the page
EB> locked around the call to updatepage.  The only real function
EB> I can see is for a data coherency lock. . .

TM> We want to serialize the case of write + write too. As I said earlier
TM> in this thread, you cannot assume that either of 2 different processes
TM> really have write permission even if the attributes seem to indicate
TM> that they do. Ugly: but that's what a stateless protocol means...

Agreed.  This is quite useful for NFS, (The caching policy in a
stateless protocol is a tradeoff discrepency/performance  tradeoff).

My focus 2.3 when hopefully all filesystems will be keeping their
dirty data in the page cache, and I don't want to pay the price for
other filesystems.

And as I have said because NFS doesn't wrap the mmap functions,
write + write isn't handled in the mmap case. 

Eric



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/