Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: I4f88d47c633096a82ef113dca2d58e584c43d1ae
Issue-ID: OOM-1744
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
|
|
This change heavily changes run_chroot.sh script.
Major change is that most of the real script logic is now run in a namespace,
so that all mounts are done in namespace.
Also, all the cleanup code and some checks are removed, also the
shell wrapper is removed.
This change is for simplifying the run_chroot.sh. It will function the same as before
from the user's point of view, except that now the script does not have to explicitly
unmount filesystems that it mounted before, so
no cleanup code isneeded. Also,
all processes run from inside a chroot should be killed immediately
when the chroot command finishes.
Change-Id: I45a3dc7c6583cb9693746cc04f42b7af325724dd
Signed-off-by: Michal Zegan <m.zegan@samsung.com>
Issue-ID: OOM-1744
|
|
In this commit I tried to follow a behaviour of docker. If mounted
path does not exist then it will handle it as a directory - it
creates it and the mountpoint is expected to be a directory also.
If the mounted path is a plain file it will do file bind mounting.
The mountpoint must be a plain file.
Change-Id: Ie8b29442c8ebb8da389ba6c9a152a49d8cf6e9be
Issue-ID: OOM-1575
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
|
|
This will change in a future commit to be little more flexible.
As of now it fixes some usage when host system has defined
some custom resolving which are needed.
Change-Id: Ifea2ba087982006c0a399d1b225d7ab13f7e94da
Issue-ID: OOM-1551
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
|
|
Script uses namespaces and overlayfs to be more robust.
Also it supports bind mounting to extend the chroot with
external files and directories.
Issue-ID: OOM-1551
Change-Id: I8c554e1b86fca86f86e6c0d0e1b106c97a08ef15
Signed-off-by: Petr Ospalý <p.ospaly@partner.samsung.com>
|