Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 20 Apr 2023 08:03:55 -0700
From:   "Darrick J. Wong" <djwong@kernel.org>
To:     Pankaj Raghav <p.raghav@samsung.com>
Cc:     Hannes Reinecke <hare@suse.de>,
        Matthew Wilcox <willy@infradead.org>,
        linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        mcgrof@kernel.org, SSDR Gost Dev <gost.dev@samsung.com>
Subject: Re: [PATCH] mm/filemap: allocate folios according to the blocksize
Message-ID: <20230420150355.GG360881@frogsfrogsfrogs>
References: <CGME20230414134914eucas1p1f0b08409dce8bc946057d0a4fa7f1601@eucas1p1.samsung.com>
 <20230414134908.103932-1-hare@suse.de>
 <2466fa23-a817-1dee-b89f-fcbeaca94a9e@samsung.com>
 <a826abe1-332f-22db-982c-ecec67a40585@suse.de>
 <1c76a3fe-5b7a-6f22-78e1-da4a54497ecd@samsung.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1c76a3fe-5b7a-6f22-78e1-da4a54497ecd@samsung.com>
Precedence: bulk

On Thu, Apr 20, 2023 at 02:28:36PM +0200, Pankaj Raghav wrote:
> On 2023-04-20 14:19, Hannes Reinecke wrote:
> >>
> >> **Questions on the future work**:
> >>
> >> As willy pointed out, we have to do this `order = mapping->host->i_blkbits - PAGE_SHIFT` in
> >> many places. Should we pursue something that willy suggested: encapsulating order in the
> >> mapping->flags as a next step?[1]
> >>
> >>
> >> [1] https://lore.kernel.org/lkml/ZDty+PQfHkrGBojn@casper.infradead.org/

I wouldn't mind XFS gaining a means to control folio sizes, personally.
At least that would make it easier to explore things like copy on write
with large weird file allocation unit sizes (2MB, 28k, etc).

> > 
> > Well ... really, not sure.
> > Yes, continue updating buffer_heads would be a logical thing as it could be done incrementally.
> >
> > But really, the end-goal should be to move away from buffer_heads for fs and mm usage. So I wonder
> > if we shouldn't rather look in that direction..
> >
> Yeah, I understand that part. Hopefully, this will be discussed as a part of LSFMM.

Agree.

--D

> 
> But the changes that are done in filemap and readahead needs to be done anyway irrespective of the
> underlying aops right? Or Am I missing something.
> 
> > Cheers,
> > 
> > Hannes
> >