[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <0da8d117-04ca-c236-dd3f-a6abbaf4097e@kernel.dk>
Date: Tue, 20 Oct 2020 08:47:39 -0600
From: Jens Axboe <axboe@...nel.dk>
To: Kent Overstreet <kent.overstreet@...il.com>
Cc: linux-kernel <linux-kernel@...r.kernel.org>,
fsdevel <linux-fsdevel@...r.kernel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Kent Overstreet <kent.overstreet@...il.com>,
Matthew Wilcox <willy@...radead.org>
Subject: Re: [PATCH 2/2] fs: generic_file_buffered_read() now uses
find_get_pages_contig
On Sat, Oct 17, 2020 at 2:11 PM Kent Overstreet <kent.overstreet@...il.com> wrote:
>
> Convert generic_file_buffered_read() to get pages to read from in
> batches, and then copy data to userspace from many pages at once - in
> particular, we now don't touch any cachelines that might be contended
> while we're in the loop to copy data to userspace.
>
> This is is a performance improvement on workloads that do buffered reads
> with large blocksizes, and a very large performance improvement if that
> file is also being accessed concurrently by different threads.
>
> On smaller reads (512 bytes), there's a very small performance
> improvement (1%, within the margin of error).
I ran this through my buffered testing, and no ill effects observed. It
also provides a nice boost on the read side for a mixed read/write
verification workload I have.
Reviewed-by: Jens Axboe <axboe@...nel.dk>
--
Jens Axboe
Powered by blists - more mailing lists