Re: [RFC] metadata updates vs. fetches (was Re: [PATCH v4] fs: Fix data race in inode_set_ctime_to_ts)

[Matthew Wilcox]

On Sun, Nov 24, 2024 at 10:24:50PM +0000, Al Viro wrote:
> On Sun, Nov 24, 2024 at 02:10:30PM -0800, Linus Torvalds wrote:
> > I *do* think that we could perhaps extend (and rename) the
> > inode->i_size_seqcount to just cover all of the core inode metadata
> > stuff.
> 
> That would bring ->i_size_seqcount to 64bit architectures and
> potentially extend the time it's being held quite a bit even
> on 32bit...

Could we just do:

again:
	nsec = READ_ONCE(inode->nsec)
	sec = READ_ONCE(inode->sec)
	if (READ_ONCE(inode->nsec) != nsec)
		goto again;

?