Messages in this thread | | | Date | Wed, 3 May 2006 18:20:29 -0400 (EDT) | From | Rik van Riel <> | Subject | Re: [RFC] kernel facilities for cache prefetching |
| |
On Tue, 2 May 2006, Linus Torvalds wrote: > On Tue, 2 May 2006, Wu Fengguang wrote: > > > > 4 files changed, 48 insertions(+), 2 deletions(-) > > So I would _seriously_ claim that the place to do all the statistics > allocation is in anything that ends up having to call "->readpage()", and > do it all on a virtual mapping level.
Why not simply read everything in a whole file at a time at boot time, while we still have enough free memory ?
We can have a small modification to the readahead code to read in the whole file on the first read or fault, or maybe even on open.
Once the system is done booting, it can switch this bootup readahead mode off through a tunable in /proc. If the system is booting on a system with not enough memory to load everything file-at-a-time, the bootup scripts can switch this off earlier (or not switch it on).
The kernel modifications needed to make this work are minimal. It might need some tweaking so we don't try to read in truly enormous files, but that is easy.
Does this sound reasonable ?
-- All Rights Reversed - 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/
| |