Re: [PATCH] drivers/staging/exfat - by default, prohibit mount of fat/vfat

From: Christoph Hellwig <hch@infradead.org>
To: "Valdis Klētnieks" <valdis.kletnieks@vt.edu>
Cc: Christoph Hellwig <hch@infradead.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Sasha Levin <alexander.levin@microsoft.com>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] drivers/staging/exfat - by default, prohibit mount of fat/vfat
Date: Mon, 2 Sep 2019 00:35:25 -0700	[thread overview]
Message-ID: <20190902073525.GA18988@infradead.org> (raw)
In-Reply-To: <295233.1567247121@turing-police>

On Sat, Aug 31, 2019 at 06:25:21AM -0400, Valdis Klētnieks wrote:
> On Fri, 30 Aug 2019 23:46:16 -0700, Christoph Hellwig said:
> 
> > Since when did Linux kernel submissions become "show me a better patch"
> > to reject something obviously bad?
> 
> Well, do you even have a *suggestion* for a better idea?  Other than "just rip
> it out"?  Keeping in mind that:

The right approach in my opinion is to start submitting patches to fs/fat
to add exfat support.  But more importantly it is to first coordinate
with other stakeholder most importantly the fs/fat/ maintainer and the
dosfstools maintainers as our local experts for fat-like file systems
instead of shooting from the hip.

> I think at that point, we can safely say that if it mounts a vfat filesystem,
> it's because the person building the kernel has gone out of their way to
> request that it do so.

Especially with boot time automounting it could happen.  Never mind that
we do not add duplicate file system drivers (or drivers in general) to
the kernel.

> Now, if what you want is "Please make it so the fat16/fat32 code is in separate
> files that aren't built unless requested", that's in fact doable and a
> reasonable request, and one that both doesn't conflict with anything other
> directions we might want to go, and also prepares the code for more easy
> separation if it's decided we really do want an exfat-only module.

No.  Assuming we even want the current codebase (which only you and
Greg seem to think so), that fat16/32 code simply has to go.

> And by the way, it seems like other filesystems rely on "others" to help out.
> Let's look at the non-merge commit for fs/ext4. And these are actual patches,
> not just reviewer comments....
> 
> (For those who don't feel like scrolling - 77.6% of the non-merge ext4 commits
> are from a total of 463 somebodies other than Ted Ts'o)
> 
> Even some guy named Christoph Hellwig gave Ted Ts'o a bunch of help.

That isn't the point.  Everyone who submitted a file system had a clear
plan where they wanted to go.  You just took someone elses out of tree
code, apparently didn't even try to understand it and are not able to
come up with what your plan for it is.  And even after repeated
questions on what that plan is duck the question and instead attack the
people asking for it.