[arch-projects] [mkinitcpio][PATCH] /dev: drop non-devtmpfs support

[Dan McGee]

On Thu, Jan 12, 2012 at 4:54 PM, Tom Gundersen <teg at jklm.no> wrote:
> This is not supported by udev any longer, so no point in keeping it around.
>
> Signed-off-by: Tom Gundersen <teg at jklm.no>
> ---
>  PKGBUILD |    4 ++--
>  init     |   14 ++------------
>  2 files changed, 4 insertions(+), 14 deletions(-)
>
> diff --git a/PKGBUILD b/PKGBUILD
> index a6a083f..7532d78 100644
> --- a/PKGBUILD
> +++ b/PKGBUILD
> @@ -8,8 +8,8 @@ license=('GPL')
>  groups=('base')
>  conflicts=('mkinitcpio')
>  provides=("mkinitcpio=$pkgver")
> -depends=('mkinitcpio-busybox>=1.16.1-2' 'module-init-tools' 'util-linux>=2.19' 'libarchive' 'coreutils'
> -         'bash' 'findutils' 'sed' 'grep' 'filesystem>=2011.10-1' 'udev>=174-1' 'file' 'gzip')
> +depends=('mkinitcpio-busybox>=1.16.1-2' 'module-init-tools' 'util-linux>=2.20.1-2' 'libarchive' 'coreutils'
> +         'bash' 'findutils' 'sed' 'grep' 'filesystem>=2011.10-1' 'udev>=177' 'file' 'gzip')
>  makedepends=('asciidoc' 'git')
>  optdepends=('xz: Use lzma or xz compression for the initramfs image'
>             'bzip2: Use bzip2 compression for the initramfs image'
> diff --git a/init b/init
> index 8cf90b1..e017ada 100644
> --- a/init
> +++ b/init
> @@ -5,18 +5,8 @@ PATH=/usr/bin
>
>  mount -t proc proc /proc -o nosuid,noexec,nodev
>  mount -t sysfs sys /sys -o nosuid,noexec,nodev
> -if mount -n -t devtmpfs udev /dev -o mode=0755,nosuid; then
> -  devtmpfs_mounted=1
> -else
> -  mount -n -t tmpfs udev /dev -o mode=0755,nosuid
> -  devtmpfs_mounted=0
> -  # We don't have devtmpfs, so add the most important standard devices
> -  mknod /dev/null c 1 3
> -  mknod /dev/zero c 1 5
> -  mknod /dev/console c 5 1
> -  # /dev/mem is needed if we want to load uvesafb before triggering uevents
> -  mknod /dev/mem c 1 1
> -fi
> +mount -t devtmpfs dev /dev -o mode=0755,nosuid
> +devtmpfs_mounted=1 # for compat only, to be removed
Compat with exactly what? It isn't like our low-level userspace stays
the same for longer than someone can blink anyways, so seems stupid to
keep this around considering other things break random hooks anyway.