ACK/Cmnt: [PATCH][B][aws] UBUNTU SAUCE: mm: swap: improve swap readahead heuristic
andrea.righi at canonical.com
Fri Dec 6 21:05:27 UTC 2019
On Fri, Dec 06, 2019 at 12:53:19PM -0800, Connor Kuehl wrote:
> On 12/3/19 2:58 AM, Andrea Righi wrote:
> > BugLink: https://bugs.launchpad.net/bugs/1831940
> > Apply a more aggressive swapin readahead policy to improve swapoff
> > performance.
> > The idea is to start with no readahead (only read one page) and linearly
> > increment the amount of readahead pages each time swapin_readahead() is
> > called, up to the maximum cluster size (defined by vm.page-cluster),
> > then go back to one page to give the disk enough time to prefetch the
> > requested pages and avoid re-requesting them multiple times.
> > Also increase the default vm.page-cluster size to 8 (that seems to work
> > better with this new heuristic).
> > Signed-off-by: Andrea Righi <andrea.righi at canonical.com>
> Was this patch was included in the AMI that received positive test results?
> I remember having a small chat about a recent round of testing for this on
> IRC but also wanted the mailing list to be informed of the positive results
> too :-)
> Acked-by: Connor Kuehl <connor.kuehl at canonical.com>
Thanks Connor, good point, I should have mentioned it in the comment.
This patch has been tested a lot and it is included in the AMI that
received positive results.
More information about the kernel-team