Messages in this thread | | | Date | Wed, 11 Jan 2006 14:42:55 -0800 | From | William Lee Irwin III <> | Subject | Re: [PATCH 1/2] hugetlb: Delay page zeroing for faulted pages |
| |
On Wed, Jan 11, 2006 at 04:02:40PM -0600, Adam Litke wrote: > I've come up with a much better idea to resolve the issue I mention > below. The attached patch changes hugetlb_no_page to allocate unzeroed > huge pages initially. For shared mappings, we wait until after > inserting the page into the page_cache succeeds before we zero it. This > has a side benefit of preventing the wasted zeroing that happened often > in the original code. The page_lock should guard against someone else > using the page before it has been zeroed (but correct me if I am wrong > here). The patch doesn't completely close the race (there is a much > smaller window without the zeroing though). The next patch should close > the race window completely.
Looks better to me.
Signed-off-by: William Irwin <wli@holomorphy.com>
-- wli - 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/
| |