Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 31 Jan 2019 11:17:24 -0500
From:   Dennis Zhou <dennis@kernel.org>
To:     David Sterba <dsterba@suse.cz>
Cc:     Dennis Zhou <dennis@kernel.org>, David Sterba <dsterba@suse.com>,
        Josef Bacik <josef@toxicpanda.com>, Chris Mason <clm@fb.com>,
        Omar Sandoval <osandov@osandov.com>,
        Nick Terrell <terrelln@fb.com>, kernel-team@fb.com,
        linux-btrfs@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 01/11] btrfs: add macros for compression type and level
Message-ID: <20190131161724.GA93554@dennisz-mbp>
References: <20190128212437.11597-1-dennis@kernel.org>
 <20190128212437.11597-2-dennis@kernel.org>
 <20190131160010.GK2900@twin.jikos.cz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20190131160010.GK2900@twin.jikos.cz>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Jan 31, 2019 at 05:00:10PM +0100, David Sterba wrote:
> On Mon, Jan 28, 2019 at 04:24:27PM -0500, Dennis Zhou wrote:
> > It is very easy to miss places that rely on a certain bitshifting for
> > decyphering the type_level overloading. Make macros handle this instead.
> 
> This encoding was a simple way to introduce the combined type and level
> for zlib and it could certainly be improved. Either we'll use helpers
> like you suggest or add a new structure that contains type and level as
> members. That way we'd see where the level still matters and where the
> just the type is ok.
> 
> I haven't checked how much intrusive this would be, so that's for later
> consideration. Some indirection can be removed for the .set_level
> callbacks as it does not necessarily need to be a function so I'm
> expecting that the code around that would be touched anyway.

The only user of type_level is btrfs_compress_pages(). I do make an
extra call to .set_level() there just to be safe, but that can be taken
out as it should be correctly set from when we mounted.

I envisioned .set_level() to be called when setting the compression
level of the mount. This can be used in the same place if we were to add
per-file compression levels too. I mention later, but an important part
of .set_level() is to repurpose the meaning of 0 from meaning use the
default to meaning use any workspace available.

Thanks,
Dennis