Re: clustering page-ins

Dr. Michael Weller (eowmob@exp-math.uni-essen.de)
Thu, 12 Aug 1999 10:35:18 +0200 (MESZ)


On Wed, 11 Aug 1999, Chuck Lever wrote:

> > so many questions :-)
>
> don't worry, i probably don't know what i'm talking about anyway... :)
>
> i'll try to post a patch soon, that will probably explain a little better.
> it's pretty simplistic, but i think it will convey my idea.

I didn't try the already posted patches, but if it is not already the case
I'd suggest to make the behaviour configurable by madvice, you know:

int madvise(void *addr, size_t len, int behav)

with behaviour like:

MADV_NORMAL The system provides no further special treatment for the
memory region.
MADV_RANDOM The system expects random page references to that memory
region.
MADV_SEQUENTIAL The system expects sequential page references to that
memory region.

the next are nice, but not as important.

MADV_WILLNEED The system expects the process will need these pages.
MADV_DONTNEED The system expects the process does not need these
pages.
MADV_SPACEAVAIL The system will ensure that memory resources are
reserved.

errno values ar EINVAL or ENOSPC for MADV_SPACEAVAIL, it return 0 or -1
for error.

I was told somehow the origins of this function might be BSD, so it might
be worthwhile to support. I saw it in many unixes but it is usually
documented to have no effect at all.

Michael.

--

Michael Weller: eowmob@exp-math.uni-essen.de, eowmob@ms.exp-math.uni-essen.de, or even mat42b@spi.power.uni-essen.de. If you encounter an eowmob account on any machine in the net, it's very likely it's me.

- 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/