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;
MIME-Version: 1.0
References: <163410.1659964655@warthog.procyon.org.uk>
In-Reply-To: <163410.1659964655@warthog.procyon.org.uk>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Mon, 8 Aug 2022 16:01:57 +0200
Message-ID: <CAK8P3a3ZNbmND-TPLEmc-37ZmK31pOT2+hDhQD+HWQZyXFZX8Q@mail.gmail.com>
Subject: Re: [RFC][PATCH] uapi: Remove the inclusion of linux/mount.h from uapi/linux/fs.h
To:     David Howells <dhowells@redhat.com>
Cc:     linux-fsdevel@vger.kernel.org, Ian Kent <raven@themaw.net>,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        Christian Brauner <christian@brauner.io>,
        linux-api@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Aug 8, 2022 at 3:17 PM David Howells <dhowells@redhat.com> wrote:
>
> Hi,
>
> We're seeing issues in autofs and xfstests whereby linux/mount.h (the UAPI
> version) as included indirectly by linux/fs.h is conflicting with
> sys/mount.h (there's a struct and an enum).
>
> Would it be possible to just remove the #include from linux/fs.h (as patch
> below) and rely on those hopefully few things that need mount flags that don't
> use the glibc header for them working around it by configuration?
>
> David
> ---
> uapi: Remove the inclusion of linux/mount.h from uapi/linux/fs.h
>
> Remove the inclusion of <linux/mount.h> from uapi/linux/fs.h as it
> interferes with definitions in sys/mount.h - but linux/fs.h is needed by
> various things where mount flags and structs are not.
>
> Note that this will likely have the side effect of causing some build
> failures.
>
> Reported-by: Ian Kent <raven@themaw.net>
> Signed-off-by: David Howells <dhowells@redhat.com>
> cc: Alexander Viro <viro@zeniv.linux.org.uk>
> cc: Christian Brauner <christian@brauner.io>
> cc: linux-fsdevel@vger.kernel.org
> cc: linux-api@vger.kernel.org

I think that's probably ok. Looking through the results from

https://codesearch.debian.net/search?q=MS_RDONLY
combined with the list of results from
https://codesearch.debian.net/search?q=linux/fs.h

I found 95 packages that mention both /somewhere/ in the code, see
below for a complete list. I did not try an exhaustive search but found
that almost all of these reference the two in different files.

The only counterexample I found is

https://sources.debian.org/src/trinity/1.9+git20200331.4d2343bd18c7b-2/syscalls/mount.c/?hl=13#L13

so this will likely break and have to get fixed to includle <linux/mount.h> or
<sys/mount.h> instead of linux/fs.h, but that is probably acceptable.
There may be others like it, but not a ton of them.

      Arnd

8<---
android-framework-23
android-platform-art
android-platform-frameworks-base
android-platform-system-core
android-platform-system-extras
android-platform-tools
apparmor
audit
avfs
bazel-bootstrap
bcachefs-tools
bpfcc
busybox
cargo
cde
ceph
chromium
criu
cryptmount
docker.io
e2fsprogs
elogind
emscripten
falcosecurity-libs
firefox
firefox-esr
fstransform
gcc-10
gcc-11
gcc-12
gcc-9
gcc-snapshot
gfarm
glibc
glusterfs
gnumach
golang-1.11
golang-1.15
golang-1.16
golang-1.17
golang-1.18
golang-1.19
golang-github-containers-storage
golang-golang-x-sys
golang-inet-netstack
hashrat
hurd
icingadb
kfreebsd-10
klibc
kubernetes
kvmtool
libblockdev
libexplain
librsvg
libvirt
linux
linux-apfs-rw
lxc
manpages-l10n
mergerfs
mozjs91
mtd-utils
network-manager
nilfs-tools
ntfs-3g
ocfs2-tools
ostree
partclone
ploop
qemu
quota
rust-libc
rustc
sash
snapd
strace
stress-ng
suricata
systemd
systemtap
testdisk
thunderbird
tiny-initramfs
tomoyo-tools
toybox
trinity
u-boot
uclibc
umview
util-linux
virtualbox
webhook
zfs-linux
zulucrypt