Re: [Cluster-devel] [PATCH v7 12/13] ext4: switch to multigrain timestamps

From: Bruno Haible <bruno@clisp.org>
To: Jan Kara <jack@suse.cz>, Xi Ruoyao <xry111@linuxfromscratch.org>,
	bug-gnulib@gnu.org
Cc: Latchesar Ionkov <lucho@ionkov.net>,
	Martin Brandenburg <martin@omnibond.com>,
	Konstantin Komarov <almaz.alexandrovich@paragon-software.com>,
	linux-xfs@vger.kernel.org, "Darrick J. Wong" <djwong@kernel.org>,
	Dominique Martinet <asmadeus@codewreck.org>,
	Christian Schoenebeck <linux_oss@crudebyte.com>,
	ecryptfs@vger.kernel.org, linux-unionfs@vger.kernel.org,
	David Howells <dhowells@redhat.com>, Chris Mason <clm@fb.com>,
	Andreas Dilger <adilger.kernel@dilger.ca>,
	Hans de Goede <hdegoede@redhat.com>,
	Marc Dionne <marc.dionne@auristor.com>,
	codalist@coda.cs.cmu.edu, linux-afs@lists.infradead.org,
	linux-mtd@lists.infradead.org,
	Mike Marshall <hubcap@omnibond.com>,
	Paulo Alcantara <pc@manguebit.com>,
	linux-cifs@vger.kernel.org,
	Eric Van Hensbergen <ericvh@kernel.org>,
	bug-gnulib@gnu.org, Miklos Szeredi <miklos@szeredi.hu>,
	Richard Weinberger <richard@nod.at>,
	Mark Fasheh <mark@fasheh.com>, Hugh Dickins <hughd@google.com>,
	Tyler Hicks <code@tyhicks.com>,
	cluster-devel@redhat.com, coda@cs.cmu.edu, linux-mm@kvack.org,
	Jeff Layton <jlayton@kernel.org>,
	Ilya Dryomov <idryomov@gmail.com>,
	Iurii Zaikin <yzaikin@google.com>,
	Namjae Jeon <linkinjeon@kernel.org>,
	Trond Myklebust <trond.myklebust@hammerspace.com>,
	Shyam Prasad N <sprasad@microsoft.com>,
	Amir Goldstein <amir73il@gmail.com>,
	Kees Cook <keescook@chromium.org>,
	ocfs2-devel@lists.linux.dev, Chao Yu <chao@kernel.org>,
	Josef Bacik <josef@toxicpanda.com>, Tom Talpey <tom@talpey.com>,
	Tejun Heo <tj@kernel.org>, Yue Hu <huyue2@coolpad.com>,
	Alexander Viro <viro@zeniv.linux.org.uk>,
	Ronnie Sahlberg <ronniesahlberg@gmail.com>,
	David Sterba <dsterba@suse.com>, Jaegeuk Kim <jaegeuk@kernel.org>,
	ceph-devel@vger.kernel.org, Xiubo Li <xiubli@redhat.com>,
	Gao Xiang <xiang@kernel.org>,
	OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>,
	Jan Harkes <jaharkes@cs.cmu.edu>,
	Christian Brauner <brauner@kernel.org>,
	linux-ext4@vger.kernel.org, Theodore Ts'o <tytso@mit.edu>,
	Joseph Qi <joseph.qi@linux.alibaba.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	v9fs@lists.linux.dev, ntfs3@lists.linux.dev,
	samba-technical@lists.samba.org, linux-kernel@vger.kernel.org,
	linux-f2fs-devel@lists.sourceforge.net,
	Steve French <sfrench@samba.org>,
	Sergey Senozhatsky <senozhatsky@chromium.org>,
	Luis Chamberlain <mcgrof@kernel.org>,
	Jeffle Xu <jefflexu@linux.alibaba.com>,
	devel@lists.orangefs.org, Anna Schumaker <anna@kernel.org>,
	Jan Kara <jack@suse.com>,
	linux-fsdevel@vger.kernel.org,
	Andrew Morton <akpm@linux-foundation.org>,
	Sungjong Seo <sj1557.seo@samsung.com>,
	linux-erofs@lists.ozlabs.org, linux-nfs@vger.kernel.org,
	linux-btrfs@vger.kernel.org, Joel Becker <jlbec@evilplan.org>
Subject: Re: [Cluster-devel] [PATCH v7 12/13] ext4: switch to multigrain timestamps
Date: Tue, 19 Sep 2023 16:52:43 +0200	[thread overview]
Message-ID: <4511209.uG2h0Jr0uP@nimes> (raw)
In-Reply-To: <1f29102c09c60661758c5376018eac43f774c462.camel@kernel.org>

Jeff Layton wrote:
> I'm not sure what we can do for this test. The nap() function is making
> an assumption that the timestamp granularity will be constant, and that
> isn't necessarily the case now.

This is only of secondary importance, because the scenario by Jan Kara
shows a much more fundamental breakage:

> > The ultimate problem is that a sequence like:
> > 
> > write(f1)
> > stat(f2)
> > write(f2)
> > stat(f2)
> > write(f1)
> > stat(f1)
> >
> > can result in f1 timestamp to be (slightly) lower than the final f2
> > timestamp because the second write to f1 didn't bother updating the
> > timestamp. That can indeed be a bit confusing to programs if they compare
> > timestamps between two files. Jeff?
> > 
> 
> Basically yes.

f1 was last written to *after* f2 was last written to. If the timestamp of f1
is then lower than the timestamp of f2, timestamps are fundamentally broken.

Many things in user-space depend on timestamps, such as build system
centered around 'make', but also 'find ... -newer ...'.

Bruno