This reverts commit da64d1a1, which
was a workaround for the inconsistent choices of interpreter path in
the early days of armhf and the limitations of copy_file at the time.
Now that copy_file automatically replicates symlinks and their
targets, we don't need to do anything special about the dynamic
linker.

An initramfs built for Debian 10 on armhf currently contains:

/lib/arm-linux-gnueabihf/ld-2.28.so           regular file
/lib/arm-linux-gnueabihf/ld-linux-armhf.so.3  link to ld-2.28.so
/lib/arm-linux-gnueabihf/ld-linux.so.3        link to /lib/ld-linux-armhf.so.3
/lib/ld-linux-armhf.so.3                      regular file

There are two *copies* of the dynamic linker as well as two symlinks
to one or the other of them.

After this commit, the initramfs contains:

/lib/arm-linux-gnueabihf/ld-2.28.so           regular file
/lib/arm-linux-gnueabihf/ld-linux-armhf.so.3  link to ld-2.28.so
/lib/ld-linux-armhf.so.3                      link to arm-linux-gnueabihf/ld-2.28.so

The /lib/…/ld-linux.so.3 symlink is no longer created, because Debian
armhf binaries all use /lib/ld-linux-armhf.so.3 as the interpreter
path, but it would be created if we ever switched over.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

This reverts commit 879602b2.  The
arm (OABI little-endian) port was replaced by armel and was last
released in Debian 5.0.  The armeb (OABI big-endian) port never made
it into a Debian release and has died off completely.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

Remove the temporary hack introduced by commit ab913066 "init:
Take over handling of rootdelay parameter from udev boot script".

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

hook-functions: preloaded libraries should not end up in the initramfs

See merge request kernel-team/initramfs-tools!19

Closes: #944929

Don't require build dependency on shellcheck when nocheck profile is enabled

See merge request kernel-team/initramfs-tools!20

hook-functions: Handle dw_mmc-rockchip hidden dependency on rockchip-io-domain

See merge request kernel-team/initramfs-tools!18

Include modules for Chromebook Plus display

See merge request kernel-team/initramfs-tools!15

Running a static checker is not a substitute for proper integration
tests, which are still TBD.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

On Samsung Chromebook Plus (v1), messages and prompts printed during an
initramfs run aren't visible to the user unless the pwm_bl, rockchipdrm,
pwm-cros-ec and panel-simple modules are included. This commit includes
these when MODULES=dep and when MODULES=most.

For MODULES=dep: adds backlight and pwm to the dep_modules_add sys-walk.
Walking backlight includes pwm_bl and pwm includes pwm-cros-ec. Since
panel-simple isn't reachable from /sys/class/**, but it can be reached
from /sys/devices/platform/edp-panel, that path is sys-walked if it
exists.

When {efi,platform,simple,vesa}-framebuffer cannot provide a working
framebuffer device, graphics is sys-walked as well. Walking drm works
for rockchipdrm too, but also includes panfrost, so graphics was chosen
instead.

For MODULES=most: the modules are included in a new case 'fb' in the
auto_add_modules intended for modules providing framebuffer support
for machines where generic drivers don't provide one.

Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>

Without rockchip-io-domain, I'm getting the following errors when
plugging micro SD cards to a rk3399-gru-kevin:

    dwmmc_rockchip fe320000.dwmmc: All phases bad!
    mmc1: tuning execution failed: -5
    mmc1: error -5 whilst initialising SD card

In that case the relevant /dev/mmcblk* device isn't available either,
so we can't boot from the external SD card.

Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>

Update to new name for kernel module therm_pm72

See merge request kernel-team/initramfs-tools!17

Also update a minor comment to clarify what to load nowadays.

See commit 6cd320996746 ("powerpc/powermac: New windfarm driver for
PowerMac G5 (AGP) and Xserve G5").

Signed-off-by: Mathieu Malaterre <malat@debian.org>

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

hook-functions: Handle spi-rockchip hidden dependency on pl330

See merge request kernel-team/initramfs-tools!16

There are still other packages with hooks and scripts that shellcheck
finds fault with.  This shouldn't cause a test failure for
initramfs-tools itself, but unfortunately that's what happens now:

https://ci.debian.net/data/autopkgtest/testing/amd64/i/initramfs-tools/2780481/log.gz



Replace the Test-Command with a script that tests only hook and boot
scripts installed by specific binary packages.  We can expand this
list as and when other packages are updated.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

Recent kernel versions (v5.0) enabled SPI DMA channels in the RK3399 DTS
file. Not including pl330 in the initramfs now makes some systems
unbootable (at least rk3399-gru-kevin) presumably because of this.

Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>

init: /dev can be noexec and /run nodev

This change *will* be visible in systemd units that use the
InaccessiblePaths option to hide device files.  systemd creates a
dummy file of each type in /run/systemd/inaccessible/<type> and
bind-mounts them over the paths specified in this option.  If a
process in such a unit attempts to access a device that is made
inaccessible, it will now fail with error EPERM and not ENODEV.
Hopefully this won't cause breakage.

See merge request kernel-team/initramfs-tools!9

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

Describe the new behaviour introduced by commit f39625af
"update-initramfs: Make "-k all" take over other initramfs images".

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

This is mnemonic, and consistent with messages the command may output.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

The fsck hook depends on logsave, which has moved from e2fsprogs
to a separate binary package.

Closes: #932854
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

"lz4cat -t" checks that the filename ends in ".lz4", not just that it
has a valid header etc.  This can be avoided by giving it the file as
stdin instead of by name.

Closes: #930366
Thanks: Dimitri John Ledkov <xnox@ubuntu.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

klibc's reboot implementation always calls reboot(2), whereas
busybox's reboot implementation defaults to signalling an init daemon
which doesn't exist in the initramfs.  The solution is simply to use
the -f option, which both implementations accept.

This seems to have been broken since commit c04a9db5 "hooks/klibc:
Make us play more nicely with busybox and static bin/sh" which caused
busybox's reboot implementation to be preferred over klibc's.

The failure to reboot was previously reported as #751488, but only
worked around by forcing a kernel panic if it failed.

Closes: #923165
Related-to: #751488
Thanks: Michael Niewöhner <linux@mniewoehner.de>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

The copy_file function applies two transformations to the target
filename:

1. If it matches /bin/*, /lib*, or /sbin/*, add /usr to the beginning
   since the initramfs is usrmerged.
2. If it refers a directory, add the basename of the source filename.

These need to be done in the opposite order, to handle a target
filename of "/bin" correctly.

Closes: #928689
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

We always need to write the resume device name to the kernel, either
(a) to resume, if a suspend image is found, or (b) to enable suspend
to that device later.  We let the kernel distinguish the two cases
itself.

Since we started reporting the attempt to resume through plymouth, if
the text front-end is used (currently the default in Debian) the
message remains on-screen and is confusing in case (b).  (If the
frame-buffer front-end is used, we clear the message.)

I can't think of a message that usefully covers both cases, and if
case (b) it's not really necessary to report anything.  Use fstype to
check for a suspend image before reporting that we're resuming.

Closes: #928736
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

mkinitramfs: suppress warning when using bzip2 or lzma

See merge request kernel-team/initramfs-tools!12

When using bzip2 a message is emitted

	W: Unknown compression command bzip2

which is wrong and irritating. Suppress this warning for both bzip2 and
lzma.

Fixes: e0d4024d ("mkinitramfs: Warn if the compression command is not recognised")
Closes: #930754

On some ChromeOS devices the keyboard is attached to an EC which the
host must communicate with over SPI.

Thanks: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

Until recently most keyboard drivers were either built-in or
HID (which we already include).  However there are other important
modular keyboard drivers such as cros_ec_keyb which were only included
when MODULES=dep.

Thanks: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

Include extcon-usbc-cros-ec for MODULES=most, and extcon drivers for MODULES=dep

See merge request kernel-team/initramfs-tools!10

Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>

Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>

/dev used to be mounted with "exec" flag due to /dev/MAKEDEV script
but that's history and it's now located in /sbin. mmap() with file
descriptor to "/dev/zero" (instead of modern mmap(,,,MAP_ANON...))
will still work.

There's no reason to have (char/block) device nodes in /run, so let's
mount it with "nodev". This does not affect sockets, fifos etc.

Signed-off-by: Topi Miettinen <toiwoton@gmail.com>

autopkgtest: Check also /usr/share/initramfs-tools/{hooks,scripts}

See merge request kernel-team/initramfs-tools!8

klibc 2.0.5-1, kmod 26-1, and systemd 240-4 fixed their shellcheck
issues. So we can also check the whole /usr/share/initramfs-tools/hooks
and /usr/share/initramfs-tools/scripts directories.

Signed-off-by: Benjamin Drung <benjamin.drung@cloud.ionos.com>

In 0.131 the automatic selection of a resume device was changed to
disable resume if the largest swap device was encrypted with a
ephemeral key.

In 0.132 I changed the automatic selection to also reject zram
devices, but I also changed it to loop over all swap devices since
zram might well be used alongside a conventional swap device that we
could use.  Unfortunately, in case swap devices are available but none
are suitable, this results in selecting the last one.

Closes: #916696
Thanks: Trek <trek00@inbox.ru>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

Related-to: #916696
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>