[PATCH v2 06/50] primitives for maintaining persisitency

[James Bottomley]

On Wed, 2025-10-29 at 05:10 +0000, Al Viro wrote:
> On Tue, Oct 28, 2025 at 08:38:00AM -0400, James Bottomley wrote:
> > On Tue, 2025-10-28 at 00:45 +0000, Al Viro wrote:
> > [...]
> > > +void d_make_discardable(struct dentry *dentry)
> > > +{
> > > +	spin_lock(&dentry->d_lock);
> > > +	dentry->d_flags &= ~DCACHE_PERSISTENT;
> > > +	dentry->d_lockref.count--;
> > > +	rcu_read_lock();
> > > +	finish_dput(dentry);
> > > +}
> > > +EXPORT_SYMBOL(d_make_discardable);
> > 
> > I was going to ask why you don't have a WARN_ON if the dentry is
> > not
> > persistent here.  Fortunately I read the next patch which gives the
> > explanation and saw that you do do this in patch 50.  For those of
> > us
> > who have a very linear way of reading and responding to patches, it
> > would have been helpful to put a comment at the top saying
> > something
> > like persistency will be checked when all callers are converted,
> > which
> > you can replace in patch 50.
> 
> Point...  How about
> void d_make_discardable(struct dentry *dentry)
> {
> 	spin_lock(&dentry->d_lock);
> 	/*
> 	 * By the end of the series we'll add 
> 	 * WARN_ON(!(dentry->d_flags & DCACHE_PERSISTENT);
> 	 * here, but while object removal is done by a few common
> helpers,
> 	 * object creation tends to be open-coded (if nothing else,
> new inode
> 	 * needs to be set up), so adding a warning from the very
> beginning 
> 	 * would make for much messier patch series.  
> 	 */
> 	dentry->d_flags &= ~DCACHE_PERSISTENT;
> 	dentry->d_lockref.count--;
> 	rcu_read_lock();  
> 	finish_dput(dentry);
> }
> 
> at that point of the series, with comment replaced with WARN_ON() in
> #50?

Ack.  Works for me.

Regards,

James