Re: [PATCH v5 08/10] xfs: Update atomic write max size

[John Garry]

On 12/03/2025 07:41, Christoph Hellwig wrote:
> On Mon, Mar 10, 2025 at 06:39:44PM +0000, John Garry wrote:
> > For RT inode, just limit to 1x block, even though larger can be supported
> > in future.
>
> Why?

Adding RT support adds even more complexity upfront, and RT is 
uncommonly used.

In addition, it will be limited to using power-of-2 rtextsize, so a 
slightly restricted feature.

> > +	if (XFS_IS_REALTIME_INODE(ip)) {
> > +		/* For now, set limit at 1x block */
>
> Why?  It' clearly obvious that you do that from the code, but comments
> are supposed to explain why something non-obvious is done.

ok

> > +		*unit_max = ip->i_mount->m_sb.sb_blocksize;
> > +	} else {
> > +		*unit_max =  min_t(unsigned int,
>
> double whitespace before the min.

will fix

> > +++ b/fs/xfs/xfs_mount.c
> > @@ -665,6 +665,32 @@ xfs_agbtree_compute_maxlevels(
> >   	levels = max(levels, mp->m_rmap_maxlevels);
> >   	mp->m_agbtree_maxlevels = max(levels, mp->m_refc_maxlevels);
> >   }
> > +static inline void
>
> Missing empty line after the previous function.

Will fix.

Thanks,
John