Re: [PATCH 1/2] ceph: conversion to new fscache API

[Jeff Layton <jlayton@kernel.org>]

On Mon, 2021-12-06 at 09:57 +0000, David Howells wrote:
> Jeff Layton <jlayton@kernel.org> wrote:
> 
> >  		if (!(gfp & __GFP_DIRECT_RECLAIM) || !(gfp & __GFP_FS))
> 
> There's a function for the first part of this:
> 
> 		if (!gfpflags_allow_blocking(gfp) || !(gfp & __GFP_FS))
> 
> > +	fsc->fscache = fscache_acquire_volume(name, NULL, 0);
> >  
> >  	if (fsc->fscache) {
> >  		ent->fscache = fsc->fscache;
> >  		list_add_tail(&ent->list, &ceph_fscache_list);
> 
> It shouldn't really be necessary to have ceph_fscache_list since
> fscache_acquire_volume() will do it's own duplicate check.  I wonder if I
> should make fscache_acquire_volume() return -EEXIST or -EBUSY rather than NULL
> in such a case and not print an error, but rather leave that to the filesystem
> to display.
> 
> That would allow you to get rid of the ceph_fscache_entry struct also, I
> think.
> 

Returning an error there sounds like a better thing to do.

I'll make the other changes you suggested now. Let me know if you change
the fscache_acquire_volume return.

> > +#define FSCACHE_USE_NEW_IO_API
> 
> That doesn't exist anymore.
> 
> > +		/*
> > +		 * If we're truncating up, then we should be able to just update
> > +		 * the existing cookie.
> > +		 */
> > +		if (size > isize)
> > +			ceph_fscache_update(inode);
> 
> Might look better to say "expanding" rather than "truncating up".
> 
> David
> 

-- 
Jeff Layton <jlayton@kernel.org>