> Hmm... It's starting to sound like we are reinventing either zero mapping,
> or lazy clear of zero pages...
nope, it's just that the original proposal suggested a much broader
solution, the malloc() cache being just one user. Also there is a subtle
difference here, zero-mmap-ing is _mandatory_ (no matter wether it's lazy
or not), while disposable pages are disposed optionally. They will not be
touched (and stay cache-hot) in 99% of the cases, and they will be
disposed and recreated in 1% of the cases. This 're-creation' is:
- in the case of the malloc cache: reallocation and zeroing
- in the case of X or web browsers: some very application
dependent (and probably costy) re-creation.
-- mingo
-
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/