| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
| |
| |
| |
| |
| |
| | |
during tmpfs install, some package scripts and triggers may
require access to /dev, /proc or /sys. make sure those are
bind mounted to have them available. they are later on move
mounted to the new root.
|
| |
| |
| |
| |
| |
| | |
they in fact are harmful, hwdrivers is not needed or even wanted
if using eudev. these were only needed for upgrade path from
alpine 1.9.x which is no longer supported.
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
we want handle all the events in queue, even if we find the devices we
are looking for. This is so we load kernel module for framebuffer even
if the apkvol and bootrepo are found first.
This means that the only valid exit condition is a poll timeout and that
we can set the timeout to 0 once we found all the things we are looking
for.
|
| |
| |
| |
| |
| |
| | |
make sure that we dont exit before the trigger thread is complete,
regardless if we find the devices/bootrepo/apkovl we are looking for or
not.
|
| | |
|
| | |
|
| |
| |
| |
| | |
we probably dont support upgrade from the old version it was there for
|
| |
| |
| |
| |
| |
| |
| | |
There are ~10k entries in /sys so we want be fast. Not all filesystems
supports dirent filed d_type, (iso9660 for example) so we need use lstat
to find directories. However, we know that sysfs supports it, so we can
avoid 10k lstat syscalls. This might be noticiable difference on rpi.
|
| | |
|
| | |
|
| |
| |
| |
| |
| | |
usb storage need some time to settle, we increase the timeout with 2
seconds if we detext usb-storage.
|
| | |
|
| | |
|
| |
| |
| |
| | |
retun success if anything of what we searched for was found.
|
| |
| |
| |
| |
| | |
avoid allocate PATH_MAX on stack for every dir. Instead reuse the path
buffer.
|
| |
| |
| |
| |
| | |
the dirent d_type is not supported on isofs apparently. Use lstat
instead.
|
| |
| |
| |
| | |
we set CLOEXEC with socket create
|
| |
| |
| |
| |
| | |
We will need set optional mount options, so we better do the mount from
script.
|
| |
| |
| |
| |
| | |
when both apkovl and boot repository are found, then we should have what
we need to setup up a working tmpfs root so we exit at that point.
|
| | |
|
| |
| |
| |
| |
| | |
add option -b OUTFILE which will create OUTFILE and add first found boot
repository
|
| |
| |
| |
| |
| | |
we need a more general recurse function so we can search for boot repos
and similar.
|
| | |
|
| | |
|
| |
| |
| |
| | |
We recursively scan /sys and trigger the events ourselves.
|
|/
|
|
|
|
|
|
| |
nlplug-findfs is a tool that will help us to find a given filesystem
using coldplug. The only reason that initramfs exist in first place is
to set up a rootfs and switch to it. The nlplug-findfs will handle
coldplug events til a given filesystem/device is found and then mount
it.
|
|
|
|
| |
mdadm and lvm needs it
|
| |
|
| |
|
|
|
|
|
| |
mdadm otherwise complains about not being able to lock
/run/mdadm/map.lock
|
| |
|
|
|
|
| |
remove bootchart and floppy from default config and add mmc
|
| |
|
| |
|
|
|
|
| |
eudev depends on devtmpfs
|
| |
|
|
|
|
|
|
| |
It breaks xen boot options like xen-pciback.hide=(01:00.0)
ref #3609
|
|
|
|
| |
and warn about root_size. Users should use rootflags instead now.
|
|
|
|
| |
we can get rid of if/else with some shell variable expansions
|
|
|
|
|
|
|
|
|
|
|
| |
Add support for two new options in the cmdline.
- rootfstype: Let's you specify the type of filesystem for the root fs
- rootflags: Let's you specify mount options for the root fs - If
specified, "ro" flag won't be added by default.
This has primarily been setup to be able to boot Alpine guests from a 9P
virtio share and make it easy to change mount options from libvirt/qemu
|
| |
|
|
|
|
|
|
|
| |
recent version of cryptsetup does not seem to work with UUID=... devices
unless udev is used to create /dev/disk/by-uuid symlinks.
We work around that by resolving UUID=... to the device.
|
|
|
|
|
|
| |
useful if wanting to optimize boot time, and do device updates
via read-only system image. package management is not too useful
in these situations.
|
|
|
|
| |
just having the fbsplash.ppm in boot media activates now splash.
|
| |
|
| |
|
| |
|
|
|
|
| |
it is useless as fbcon is really loaded by mdev.
|