buildroot/system/skeleton, branch 2019.02-op-build

skeleton: use BR2_SYSTEM_DEFAULT_PATH as default PATH

2018-12-31T13:32:44+00:00

We substitute the path specified in system/skeleton/etc/profile with the path specified in the configuration variable $(BR2_SYSTEM_DEFAULT_PATH). $(BR2_SYSTEM_DEFAULT_PATH) is a Kconfig string, so it is already double quoted. This means that export PATH=value will now be export PATH="value" in /etc/profile, which is perfectly fine. Signed-off-by: Markus Mayer Reviewed-by: "Yann E. MORIN" [Thomas: rework commit log about the double quoting] Signed-off-by: Thomas Petazzoni

skeleton: PAGER without blank and unset at end of for loop

2018-06-05T16:50:49+00:00

The PAGER environment variable is including a blank character at the end. Remove this. A for loop has been unsetting the variable inside the loop, this is only needed once at the end of the loop. Signed-off-by: Florian La Roche Signed-off-by: Thomas Petazzoni

skeleton: add /dev/fd, /dev/std{in, out, err} symlinks for static /dev on readonly rootfs

2018-05-01T19:53:45+00:00

Some applications, e.g. bashs process subsitution feature, rely on the convention of `/dev/fd` being a symbolic link to `/proc/self/fd`. When a static /dev is used on a readonly rootfs then the runtime ln invocations in the inittab will fail, so we need to add the symlinks at build time. Makedevs doesn't support creating symlinks, so instead add the symlinks to the default skeleton. For non-static /dev setups, the kernel will mount devtmpfs which shadows the /dev of the rootfs, but then the runtime ln invocations in inittab will create the symlinks. Signed-off-by: Peter Korsgaard

system: separate sysv and systemd parts of the skeleton

2017-08-02T18:04:29+00:00

For systemd, we create a simple /etc/fstab with only an entry for /, as systemd otherwise automatically mounts what it needs where it needs it. systemd does not like that the content of /var be symlinks to /tmp, especially journald that starts before /tmp is mounted, and thus the journal files are hidden from view, which causes quite a bit of fuss... Instead, move the current /var to a sysv-only skeleton. systemd at install time will create the /var content it needs, so we just create an empty /var for systemd. systemd would create /home and /srv at runtime if they are missing, but it is better to create them right now, to simplify supporting systemd on a RO filesystem in the (near) future. Signed-off-by: "Yann E. MORIN" Cc: Maxime Hadjinlian Reviewed-by: Arnout Vandecappelle (Essensium/Mind) Signed-off-by: Arnout Vandecappelle (Essensium/Mind) Signed-off-by: Thomas Petazzoni

package/ifupdown-scripts: new package

2017-07-04T21:38:18+00:00

The ifupdown scripts can be used independently of the init system, be it sysv, busybox or systemd; they could even be used when there is no init system (i.e. the user is providing his own). Currently, those ifupdown scripts are bundled in the skeleton. But we soon will have a skeleton specific to systemd, so we would be missing those scripts (when systemd-networkd is not enabled). So, move those scripts to their own package. To keep the current behaviour (before it is changed in future commits), we make that package default to y, but depend on the default skeleton. Instead of being a target-finalize hook, the scripts are installed as any other package are, with a package install-target command. Signed-off-by: "Yann E. MORIN" Cc: Arnout Vandecappelle Cc: Thomas Petazzoni Cc: Maxime Hadjinlian Reviewed-by: Arnout Vandecappelle (Essensium/Mind) [Thomas: drop empty IFUPDOWN_SCRIPTS_SOURCE] Signed-off-by: Thomas Petazzoni

skeleton: fix permissions on /dev/pts/ptmx

2017-03-18T14:24:56+00:00

Without this patch, it is not possible to allocate PTYs when a generated rootfs image with a recent glibc and systemd is launched as a container on an RHEL7 system via machinectl/systemd-nspawn. The container boots, but `machinectl login mycontainer` fails. The culprit is /dev/pts/ptmx with 0000 perms. On a typical system, there are two `ptmx` devices. One is provided by the devpts at /dev/pts/ptmx and it is typically not directly accessed from userspace. The other one which actually *is* opened by processes is /dev/ptmx. Kernel's documentation says these days that /dev/ptmx should be either a symlink, or a bind mount of the /dev/pts/ptmx from devpts. When a container is launched via machinectl/machined/systemd-nspawn, the container manager prepares a root filesystem so that the container can live in an appropriate namespace (this is similar to what initramfs is doing on x86 desktops). During these preparations, systemd-nspawn mounts a devpts instance using a correct ptmxmode=0666 within the container-to-be's /dev/pts, and it adds a compatibility symlink at /dev/ptmx. However, once systemd takes over as an init in the container, /lib/systemd/systemd-remount-fs applies mount options from /etc/fstab to all fileystems. Because the buildroot's template used to not include the ptmxmode=... option, a default value of 0000 was taking an effect which in turn led to not being able to allocate any pseudo-terminals. The relevant kernel option was introduced upstream in commit 1f8f1e29 back in 2009. The oldest linux-headers referenced from buildroot's config is 3.0, and that version definitely has that commit. Mount options that are not understood by the system are anyway ignored, so backward compatibility is preserved. Signed-off-by: Jan Kundrát Reviewed-by: Arnout Vandecappelle (Essensium/Mind) [Thomas: fix commit title, adjust commit log.] Signed-off-by: Thomas Petazzoni

system/skeleton: use uid/gid 65534 for nobody/nogroup

2016-08-26T13:39:42+00:00

As recently discussed on lwn.net: https://lwn.net/Articles/695478/ The kernel has special behaviour for uid/gid 65534: 1. The kernel maps UIDs > 65535 to it when some subsystem/API/fs only supports 16bit UIDs, but a 32bit UID is passed to it. 2. it's used by the kernel's user namespacing as the internal UID that external UIDs are mapped to that don't have any local mapping. 3. It's used by NFS for all user IDs that cannot be mapped locally if UID mapping is enabled. Most distributions already map (or are in the progress of changing) nobody/nogroup to the 65534 uid/gid, so lets do so as well. Signed-off-by: Peter Korsgaard Acked-by: Arnout Vandecappelle (Essensium/Mind) Signed-off-by: Thomas Petazzoni

package/skeleton: remove useless .empty file

2016-07-05T13:57:20+00:00

We now have a real file in that directory, so we do not need a .empty placeholder. Signed-off-by: "Yann E. MORIN" Signed-off-by: Thomas Petazzoni

system/skeleton: update etc/mtab with a more sensible link

2016-07-05T13:24:38+00:00

Currently, our /etc/mtab points to /proc/mounts. This was all neat so far, and was good for a sysv-like init system. However, the way today is to point it at /proc/self/mounts, the per-process mount tab. Additionnally, that's what systemd expects. If /etc/mtab is not a symlink to ../proc/self/mounts and the rootfs is readonly, systemd would whine loudly (and a service unit would be marked failed). Since it works well for sysv-like init systems too, just use that. Signed-off-by: "Yann E. MORIN" Reviewed-by: Romain Naour Reviewed-by: Arnout Vandecappelle (Essensium/Mind) Signed-off-by: Thomas Petazzoni

skeleton: Recreate /var/run symlink

2016-02-11T22:20:38+00:00

It was the only one with a trailing slash, it's only for coherence's sake. Signed-off-by: Maxime Hadjinlian Acked-by: Arnout Vandecappelle (Essensium/Mind) Signed-off-by: Peter Korsgaard