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:   Wed, 19 Jul 2023 13:51:25 -0700
From:   Dmitry Torokhov <dmitry.torokhov@gmail.com>
To:     Christoph Hellwig <hch@infradead.org>
Cc:     Christian Brauner <brauner@kernel.org>,
        Guenter Roeck <linux@roeck-us.net>,
        Bill O'Donnell <billodo@redhat.com>,
        Rob Barnes <robbarnes@google.com>, bleung@chromium.org,
        linux-fsdevel@vger.kernel.org,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] fs: export emergency_sync
Message-ID: <ZLhMzQWUS0htHEdb@google.com>
References: <20230718214540.1.I763efc30c57dcc0284d81f704ef581cded8960c8@changeid>
 <ZLcOcr6N+Ty59rBD@redhat.com>
 <ad539fad-999b-46cd-9372-a196469b4631@roeck-us.net>
 <20230719-zwinkert-raddampfer-6f11fdc0cf8f@brauner>
 <ZLeDVcQrFft8FYle@infradead.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <ZLeDVcQrFft8FYle@infradead.org>
Precedence: bulk

On Tue, Jul 18, 2023 at 11:31:49PM -0700, Christoph Hellwig wrote:
> On Wed, Jul 19, 2023 at 07:53:32AM +0200, Christian Brauner wrote:
> > On vacation until next. Please add a proper rationale why and who this
> > export is needed by in the commit message. As right now it looks like
> > someone thought it would be good to have which is not enough for
> > something to become an export.
> 
> emergency_sync is a relaly bad idea and has all kinds of issues.
> It should go away and not grow more users outside of core code,
> and the one Guenther points to should never have been added.
> 
> If we want to allow emergency shutdowns it needs a proper interface
> and not a remount read-only ignoring some rules that tends to make
> things worse and instad of better, and even for that I'm not sure
> I want modules to be able to drive it.

I am not sure why you would not want modules to use it - in the case we
have here we detect a catastrophic failure in a critical system
component (embedded controller crashed) and would like to have as much
of the logs saved as possible. It is a module because this kind of EC
may not be present on every system, but when it is present it is very
much a core component.

Thanks.

-- 
Dmitry