On 6/15/16, 03:08, "linux-nfs-owner@xxxxxxxxxxxxxxx on behalf of Christoph Hellwig" <linux-nfs-owner@xxxxxxxxxxxxxxx on behalf of hch@xxxxxxxxxxxxx> wrote:
>On Tue, Jun 14, 2016 at 03:05:09PM -0400, Trond Myklebust wrote:
>> Commits are no longer required to be serialised.
>
>But we need something to lock out new callers of inode_dio_end
>when calling inode_dio_wait. Then again the current code already
>fails to to do that..
We could do that by setting the I/O mode to buffered; that still allows parallelism for the buffered I/O case. I wouldn’t normally expect applications that use O_DIRECT on NFS to call fsync(), since we always enforce O_DIRECT|O_SYNC semantics.
��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
