On Tue, May 16, 2000 at 11:57:25PM +0200, iafilius@xs4all.nl wrote:
> Just noticed while copying a (large/tested with 60MB) file it takes about
> twice the file's size in cache (which takes it from free memory).
>
> This must be because kernel doesn't know about the same files,
> and caches them separate.
They are separate files after all, so what should it do? Do a memcmp
over all of its caches to find whether there are identical cache
pages?
> But it _is_ waste of memory isn't ?
There were proposals for interfaces so that applications can give
hints to the kernel on how they use the files to let the kernel make
better decisions on caching. I don't know what of that is already
working somewhere.
And those hints could always make it worse. Take your copy for
example: cp tells the kernel it will only stream write the other file,
then the kernel makes no effort to keep the written pages in RAM.
Immediately after cp terminates another process accesses the copied
file, which now has to be read in again even though it was there a few
milliseconds earlier anyway.
-- Andreas E. Bombe <andreas.bombe@munich.netsurf.de> DSA key 0x04880A44 http://home.pages.de/~andreas.bombe/ http://linux1394.sourceforge.net/- 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/
This archive was generated by hypermail 2b29 : Tue May 23 2000 - 21:00:13 EST