[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20131121152453.056a6417df2f342d898864ad@canb.auug.org.au>
Date: Thu, 21 Nov 2013 15:24:53 +1100
From: Stephen Rothwell <sfr@...b.auug.org.au>
To: Dave Kleikamp <dave.kleikamp@...cle.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
Christoph Hellwig <hch@...radead.org>,
LKML <linux-kernel@...r.kernel.org>,
"linux-fsdevel@...r.kernel.org" <linux-fsdevel@...r.kernel.org>,
"Maxim V. Patlasov" <mpatlasov@...allels.com>, linux-aio@...ck.org,
Kent Overstreet <kmo@...erainc.com>,
Jens Axboe <axboe@...nel.dk>
Subject: Re: [GIT PULL] direct IO support for loop driver
Hi all,
On Wed, 20 Nov 2013 16:46:38 -0600 Dave Kleikamp <dave.kleikamp@...cle.com> wrote:
>
> Yeah, I'll work on something cleaner. The vital piece is some container
> that can either contain an iovec or or an array of kernel pages to be
> passed to the filesystems. The iov_iter structure was a reasonable
> choice, but the implementation is a bit sloppy. Kent wants to pass the
> bio structure itself, while this patchset used the biovec array. As long
> as I'm reworking the patchset, I'll see if I can accommodate his
> requirements.
So given this rework, I think I should drop the aio-direct tree from
linux-next for now. We can add it back (hopefully with a better name)
when the rewrite is in a reasonable state.
--
Cheers,
Stephen Rothwell sfr@...b.auug.org.au
Content of type "application/pgp-signature" skipped
Powered by blists - more mailing lists