From: Jan Kara Subject: Re: [PATCH 17/19] ext4: Add support for MAP_SYNC flag Date: Thu, 12 Oct 2017 15:42:09 +0200 Message-ID: <20171012134209.GE29293@quack2.suse.cz> References: <20171011200603.27442-1-jack@suse.cz> <20171011200603.27442-18-jack@suse.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , linux-fsdevel , linux-ext4 , linux-xfs@vger.kernel.org, Christoph Hellwig , Ross Zwisler , Ted Tso , "Darrick J. Wong" To: Dan Williams Return-path: Received: from mx2.suse.de ([195.135.220.15]:38147 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751004AbdJLNmM (ORCPT ); Thu, 12 Oct 2017 09:42:12 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: On Wed 11-10-17 15:11:21, Dan Williams wrote: > On Wed, Oct 11, 2017 at 1:06 PM, Jan Kara wrote: > > Now when everything is prepared, add support in ext4 to accept MAP_SYNC > > as an mmap(2) flag. > > > > Signed-off-by: Jan Kara > > --- > > fs/ext4/file.c | 20 ++++++++++++++++++++ > > 1 file changed, 20 insertions(+) > > > > diff --git a/fs/ext4/file.c b/fs/ext4/file.c > > index 61a8788168f3..f013cda84b3d 100644 > > --- a/fs/ext4/file.c > > +++ b/fs/ext4/file.c > > @@ -364,6 +364,25 @@ static int ext4_file_mmap(struct file *file, struct vm_area_struct *vma) > > return 0; > > } > > > > +#define EXT4_SUPPORTED_MAP_FLAGS (LEGACY_MAP_MASK | MAP_SYNC) > > + > > +static int ext4_file_mmap_validate(struct file *file, > > + struct vm_area_struct *vma, > > + unsigned long map_flags) > > +{ > > + if (map_flags & ~EXT4_SUPPORTED_MAP_FLAGS) > > + return -EOPNOTSUPP; > > + > > + /* > > + * We don't support synchronous mappings for non-DAX files. At least > > + * until someone comes with a sensible use case. > > + */ > > + if (!IS_DAX(file_inode(file)) && (map_flags & MAP_SYNC)) > > + return -EOPNOTSUPP; > > Perhaps EPERM instead to differentiate the unsupported flags case? > There's precedent for mmap returning EPERM for reasons other than > incompatible PROT flags. Hum, I could make it EINVAL. EPERM looks just too bogus to me. Honza -- Jan Kara SUSE Labs, CR