[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210813070231.GA26339@lst.de>
Date: Fri, 13 Aug 2021 09:02:31 +0200
From: Christoph Hellwig <hch@....de>
To: Matthew Wilcox <willy@...radead.org>
Cc: David Howells <dhowells@...hat.com>,
"Darrick J. Wong" <darrick.wong@...cle.com>,
Christoph Hellwig <hch@....de>,
Trond Myklebust <trond.myklebust@...marydata.com>,
linux-nfs@...r.kernel.org, viro@...iv.linux.org.uk,
jlayton@...nel.org, sfrench@...ba.org,
torvalds@...ux-foundation.org, linux-mm@...ck.org,
linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH v2 1/5] nfs: Fix write to swapfile failure due to
generic_write_checks()
On Fri, Aug 13, 2021 at 04:09:39AM +0100, Matthew Wilcox wrote:
> On Thu, Aug 12, 2021 at 09:22:06PM +0100, David Howells wrote:
> > Trying to use a swapfile on NFS results in every DIO write failing with
> > ETXTBSY because generic_write_checks(), as called by nfs_direct_write()
> > from nfs_direct_IO(), forbids writes to swapfiles.
>
> Why does nfs_direct_write() call generic_write_checks()?
>
> ie call generic_write_checks() earlier, and only swap would bypass them.
Yes, something like that is a good idea probably. Additionally I'd like
to move to a separate of for swap I/O ASAP given that NFS only
implemens ->direct_IO for swap.
Powered by blists - more mailing lists