Re: please revert bogus patch to vmscan.c

From: Linus Torvalds (torvalds@transmeta.com)
Date: Mon Oct 29 2001 - 18:25:36 EST


On Mon, 29 Oct 2001, Benjamin LaHaise wrote:
>
> I think it's far more expensive to pull a page back in from disk.

Note that _if_ the page is so often accessed that it stays in the TLB, it
won't ever hit the disk. We _will_ do a TLB invalidate when we unmap it,
and it will get re-mapped long before it's written out.

So it's more likely to result in extra soft-faults, but considering that
if the system is paging stuff out we'll almost certainly be doing enough
context switches that the thing doesn't matter.

And avoiding the extra TLB flush (for _every_ page scanned) is noticeable
according to Andrea.

                Linus

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



This archive was generated by hypermail 2b29 : Wed Oct 31 2001 - 21:00:38 EST