Date:   Thu, 16 Dec 2021 14:12:58 +0100
From:   =?utf-8?B?TWljaGHFgiBLxJlwaWXFhA==?= <kernel@kempniu.pl>
To:     Boris Brezillon <boris.brezillon@collabora.com>
Cc:     Miquel Raynal <miquel.raynal@bootlin.com>,
        Richard Weinberger <richard@nod.at>,
        Vignesh Raghavendra <vigneshr@ti.com>,
        linux-mtd@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 0/5] mtdchar: add MEMREAD ioctl
Message-ID: <Ybs7WrAQTat5Gsjg@larwa.hq.kempniu.pl>
References: <20211216083418.13512-1-kernel@kempniu.pl>
 <20211216112456.327298dd@collabora.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20211216112456.327298dd@collabora.com>
Precedence: bulk

> Splitting patch 1 and 5 means you have an incompatible ABI change
> between those 2 commits, thus breaking bisectability. I'd recommend
> putting patches 2-4 first and squashing patch 1 and 5 in a single
> commit placed at the end of the series. The other options would be to
> add a way to extend ioctls in a backward compatible way (the DRM
> subsystem does that by filling the unspecified part of the struct with
> zeros, and relying on the fact that 0 values always implies 'default
> behavior' when the struct is extended [1]).

Ack, thanks.  I will follow your recommendation in v3 and in the
meantime the existing split should make it easier to follow how v2
extends v1 with ECC statistics handling.

-- 
Best regards,
Michał Kępień