At run time, /run and /tmp get overmounted with a tmpfs, so anything
that is there becomes inaccessible.
Scripts in the fakeroot environment could call tools preparing the early
environment, leaving traces in /run or /tmp. For example, mkusers might
create home directories in /run: openssh sets the sshd home directory to
/run/sshd, so mkusers creates it. But since a tmpfs is mounted over it,
it doesn't exist at runtime, so the openssh service creates it at
startup (and deletes it when the service is stopped).
In addition, packages or rootfs overlay may leave things there as well.
Those may actually pose a runtime problem because the created file or
directory is missing - or it may not be a problem because the package
creates the missing files/directories on startup. In this situation,
it's better not to have them in the rootfs image (because they're not
functional anyway), but it's good to leave them in TARGET_DIR to make it
easier to debug the situation.
Therefore, remove the contents of /run and /tmp in the fakeroot
environment after ROOTFS_PRE_GEN_HOOKS, so they are still left in
TARGET_DIR.
Signed-off-by: Norbert Lange <nolange79@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
U-Boot mkimage zstd is available since v2020.10:
26073f9ed3
Signed-off-by: José Luis Salvador Rufo <salvador.joseluis@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The fakeroot script does not appear to be used in any of the checked
in defconfig targets, but it seems that most often the post
fakeroot script should be done after all the packages rules have
been applied instead of before.
Given that a change in systemd moved the SYSTEMD_PRESET_ALL hook to
a ROOTFS_PRE_CMD_HOOKS, there was no way to use a FAKEROOT script
to disable a service or fixup a systemd configuration. The systemd
move makes sense, and this just tries to preserve the same ability
to fixup a rootfs after all the cmd hooks are processed.
Refer to commit 65b63785a6 for
the change that instigated this reordering.
Signed-off-by: Charles Hardin <ckhardin@gmail.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
This patch is cosmetic and moves down ROOTFS_REPRODUCIBLE for
consistency.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Set the SELinux file security contexts using setfiles when generating
root filesystem images.
Without such security contexts created at build time, they need to be
setup at first boot by running the restorecon utility on the target.
This has two drawbacks:
- You have to special case the first boot, which cannot be done in
enforcing mode, and will have to run restorecon, then reboot.
- You cannot support read-only filesystems.
By setting up the security contexts at build time, we can have a
filesystem image that is immediately ready to boot an SELinux system
in enforcing mode, including if the root filesystem is read-only.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
xz help indicates only 1 thread is used unless we set threads:
-T, --threads=NUM use at most NUM threads; the default is 1; set to 0
to use as many threads as there are processor cores
Since this splits the file into blocks, the result will be not
bit-for-bit identical to single-threaded compression. Therefore, don't
enable this in BR2_REPRODUCIBLE builds.
Signed-off-by: James Hilliard <james.hilliard1@gmail.com>
Reviewed-by: Matthew Weber <matthew.weber@rockwellcollins.com>
[Arnout: append the option instead of repeating the entire command]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Using the GZIP environment variable to pass gzip options is
deprecated, and therefore we are going to remove the "GZIP = -n"
definition from the main Buildroot Makefile. In preparation for this,
we explicitly add the -n argument to the gzip call in fs/common.mk to
ensure reproducibility.
Signed-off-by: Atharva Lele <itsatharva@gmail.com>
Acked-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
show-dependency-tree was introduced in this release cycle, as a way to
quickly and easily provide the dependency tree to graph-depends.
show-dependency-tree is no longer used, now that graph-depends has been
switched over to using the more versatile show-info.
Beside, show-dependency-tree has never been part of a release.
Drop it.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Sometimes, it is need to quickly get the metadata of a subset of
packages, without resorting to a full-blown JSON query.
Introduce a new per-package (and per-filesystem) foo-show-info rule,
that otputs a per-entity valid JSON blob.
Note that calling it for multiple packages and.or filesystems at once
will not generate a valid JSON blob, as there would be no separator
between the JSON elements:
$ make {foo,bar}-show-info
{ "foo": { foo stuff } }
{ "bar": { bar stuff } }
However, jq is able to absorb this, with its slurping ability, which
generates an array (ellipsed and manualy reformated for readability):
$ make {foo,bar}-show-info |jq -s . -
[
{ "foo": { foo stuff } },
{ "bar": { bar stuff } }
]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Currently, only first-level dependencies of a filesystem are added to
the global list of packages, thus missing all recursive dependencies.
Use the newly introduced recursive variable instead, which already
contains the rootfs-common dependencies too.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
This allows getting all the recursive dependencies of filesystems,
ike we have for packages, and allows us to treat both in a similar
fashion.
Reported-by: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
This makes the filesystems resemble packages yet a bit more, and will
allow sorting "items" on their type and names, when indexed from the
upper-case names.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
fakeroot by default forwards {f,l,}chown calls to libc and ignores
permission issues, which may cause issues when building in restricted
environments like user namespaces as set up with bubblewrap where a chown
call with a uid/gid not mapped in the user namespace instead returns EINVAL.
This error is not masked by fakeroot and returned to the caller, causing
failures.
There is no real reason to really perform the *chown calls in the context of
Buildroot (as the calls will likely just fail and files are not accessed
outside the fakeroot environment any way).
This forwarding can be disabled by setting the FAKEROOTDONTTRYCHOWN
environment variable, so set it when fakeroot is executed.
Reported-by: Esben Nielsen <nielsen.esben@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, when we need to build the full dependency graph, we call make
to show the list of packages (make show-targets), and then call it again
and again iteratively while it returns new packages.
Since calling make will parse the whole set of our Makefiles, this takes
quite a bit of time (~4s each here), and the total can get pretty long.
However, make being make, already builds the whole dependency tree
information, so we can just ask for it.
Add a new top-level rule 'show-dependency-tree' that displays the whole
set of dependencies for all packages. For each package, its name, type
and version is displayed, then all the direct, first-level dependencies
are dumped. We choose a format that is not unlike the dot-graph format,
because it is both easy to read as a human, and easy to parse as a
machine:
foo: target 1.2.3
foo -> bar host-meh
bar: target virtual
bar -> buz
buz: target 2.3.4
buz ->
host-meh: host virtual
host-meh -> host-bleark
host-bleark: host 3.4.5
host-bleark ->
rootfs-meh: host
rootfs-meh -> host-bleark
To be noted: rootfs are currently reported as if they were 'host'
packages, to stay aligned with how graph-depends currently treats them.
Ideally, graph-depends could be enhanced to recognise them separately,
but that is another story.
For just plain defconfig, which is about the smallest config we can have
with an internal toolchain, we already have a seven-fold improvement
(with the graph-depends rule modified to not run the pdf generation, to
be able to just compare the tree generation):
$ time make graph-depends
real 0m27.344s
$ time make show-dependency-tree
real 0m3.848s
>From defconfig, C++, wchar, locales, ssp, and allyespackageconfig,
tweaked for even more packages (qt5 not qt4, luajit to avoid multi
providers, etc...), the timings are (graph-depends still modified to
not generate the pdf):
$ time make graph-depends
real 1m56.459s
$ time make show-dependency-tree
real 0m5.748s
There. I don't think those numbers need any explanation whatsoever;
they do speak on their own. OK, for maths sake, the ratio is about
twenty-fold. So, "yeah", I guess... ;-)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
The static devices defined by packages are currently added to the full
device table when two conditions are met:
(1) ROOTFS_DEVICE_TABLES is non-empty
(2) BR2_ROOTFS_DEVICE_CREATION_STATIC=y
(2) is obviously correct. However, depending on (1) is not correct: if
the user doesn't provide any custom permission table and custom device
table, then ROOTFS_DEVICE_TABLES will be empty.
So instead, move the addition of the package-defined static devices
outside of condition (1), and have it only under condition (2).
Reported-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
In commit 6b50f988ad ("fs/common.mk:
rename internal variable"), USERS_TABLE was renamed to
ROOTFS_FULL_USERS_TABLE.
This commit follows the same direction by renaming the
FULL_DEVICE_TABLE variable to ROOTFS_FULL_DEVICE_TABLE.
In addition, for consistency, the file itself is renamed
full_device_table.txt.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
In preparation of more renames, rename the variable that points to the
final users table.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
[Thomas: as suggested by Arnout, use ROOTFS_FULL_USERS_TABLE instead
of ROOTFS_FINAL_USERS_TABLE.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Some filesystems may want to tweak their output names, rather than using
the fixed "rootfs.foo" scheme. Add a ROOTFS_FOO_IMAGE_NAME variable for
this purpose.
Signed-off-by: Carlos Santos <casantos@datacom.com.br>
[yann.morin.1998@free.fr: fix the patch]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
In the current code, the creation of the main output directories
(BUILD_DIR, STAGING_DIR, HOST_DIR, TARGET_DIR, etc.) is done by a
global "dirs" target. While this works fine in the current situation,
it doesn't work well in a context where per-package host and target
directories are used.
For example, with the current code and per-package host directories,
the output/staging symbolic link ends up being created as a link to
the per-package package sysroot directory of the first package being
built, instead of the global sysroot.
This commit reworks the creation of those directories by having the
package/pkg-generic.mk code ensure that the build directory, target
directory, host directory, staging directory and binaries directory
exist before they are needed.
Two new targets, host-finalize and staging-finalize are added in the
main Makefile to create the compatibility symlinks for host and
staging directories. They will be extended later with additional logic
for per-package directories.
Thanks to those changes, the global "dirs" target is entirely removed.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The command "make show-build-order" doesn't show dependencies of rootfs-common target.
This patch adds $(ROOTFS_COMMON_DEPENDENCIES) to PACKAGES variable.
Signed-off-by: Serj Kalichev <serj.kalichev@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Since 118534fe54 (fs: use a common tarball as base for the other
filesystems), the filesystem creation is split in two steps, using an
intermediate tarball to carry the generic, common finalisations to the
per-filesystem finalisation and image creation.
However, this intermediate tarball causes an issue with capabilities:
they are entirely missing in the generated filesystems.
Capabilities are stored in the extended attribute security.capability,
which tar by default will not store/restore, unless explicitly told to,
e.g. with --xattrs-include='*', which we don't pass.
Now, passing this option when creating and extracting the intermediate
tarball, both done under fakeroot, will cause fakeroot to report an
invalid filetype for files with capabilities. mksquashfs would report
such unknown files as a warning, while mkfs.ext2 would fail (with a
similar error message), e.g.:
File [...]/usr/sbin/getcap has unrecognised filetype 0, ignoring
This is due to a poor interaction between tar and fakeroot; running as
root the exact same commands we run under fakeroot, works as expected.
Unfortunately, short of fixing fakeroot (which would first require
understanding the problem in there), we don't have much options.
The intermediate tarball was made to avoid redoing the same actions over
and over again for each filesystem to build. However, most of the time,
only one or two such filesystems would be enabled [0], and those actions
are usually pretty lightweight. So, using an intermediate tarball does
not provide a big optimisation.
The main reason to introduce the intermediate tarball, however, is that
it allows to postpone per-filesystem finalisations to be applied only
for the corresponding filesystem, not for all of them.
So, we get rid of the intermediate tarball, and simply move all of the
code to run under fakeroot to the per-filesystem fakeroot script.
Instead of extracting the intermediate tarball, we just rsync the
original target/ directory, and apply the filesystem finalisations on
that copy. The only thing still done in the rootfs-common step is to
generate the intermediate files (users file, devices file) that are used
in the fakeroot script.
Fixes: https://bugs.busybox.net/show_bug.cgi?id=11216
Note: an alternate solution would have been to keep the intermediate
tarball to keep most of the common finalisations, and move only the
permissions to each filesystem, but that was getting a bit more complex
and changed the ordering of permissions and post-fakeroot scripts. Once
we bite the bullet of having some common finalisation done in each
filesystem, it's easier to just move all of them.
[0] Most probsably, users would enable the real filesystem to put on
their device, plus the 'tar' filesystem, to be able to easily inspect
the content on their development machine.
Reported-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Currently, the filesystems do not depend on building host-tar when it
is needed, even though all of them have to extract the intermediate
tarball.
However, in degenerate (but legally valid) configurations with no
user-selectable package selected, host-tar would not be built, so the
rootfs images would use whatever improper tar the system has.
Add the conditional dependency to host-tar to the rootfs-common
intermediate image. Since this is the internal step that all real rootfs
generators depend on, they now properly depend on host-tar when needed.
In practice, when host-tar is needed, it will always be built before the
rootfs images, because it is a dependency of all packages (except a very
few, like the skeleton), of which host-fakeroot, which is a mandatory
dependency of rootfs-comon anyway. But for consistency sake, let's
explicitly add host-tar as a dependency to rootfs-common too.
Note that rootfs-tar already had that dependency, and we leave it as-is
because it is semantically correct, even if superfluous.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Currently, when a custom user table and a package define the same user,
the settings from the package takes precedence over the ones from the
custom user table.
However, it makes sense to allow the settings from the custom user table
take precedence. For example, it would allow redirecting the user's
home directory to an alternate location (e.g. away from tmp and into a
partition that is persistent).
The support/scripts/mkusers script will only retain settings from the
latest definition it finds.
Thus, by passing the custom user table after the package defined users,
it is possible to override the package provided user definitions.
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Matthew Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Fixes#11046
Commit bb2a57a17a (fs: run packages' filesystem hooks in a copy of target/)
changed the file system logic to run file system hooks from packages on a
copy of TARGET_DIR, and finally use this copy as input for the file system
generation.
This copy was done with rsync, which by default does not preserve hard
links, leading to an expansion of the file system images when hard links are
present.
Fix it by passing the -H option to rsync (preserve hard links).
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Each of the intermediate, per-rootfs target directories, as well as the
intermediate tarball, can take quite some place, and is mostly a
duplication of what's already in target/. The only delta, if any, would
be the tweaks made by the filesystem image generations, but those tweaks
are most probably only meaningful when seen as root.
We normally do not remove intermediate files, but those can be quite
large, and are not directly usable by, nor accessible to the user.
So, get rid of them once the filesystem has been generated.
This does not need to be done in fakeroot.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Tested-by: Matt Weber <matthew.weber@rockwellcollins.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Now that the pre-fs ones are run on a transient copy of target/, the
post-fs hooks are no longer needed because we no longer need to restore
the target/ directory as it is only a internal copy.
Remove support for the post-fs hooks, and update the sole package using
them.
We do not add a legacy check because this was mostly a purely-internal
detail that was never really exposed nor documented.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, some packages may register hooks to be run just before and
just after the generic tarball image is generated, because they need to
prepare the filesystem for read-only or read-write operation.
However, this means that, if any of the hooks or the image generation
fails, the target directory is left in a dangling, inconsistent state.
We fix that by doing a copy of target/, run the hooks on that copy,
generate the generic tarball image out of that, and get rid of the copy.
This way, we can guarantee consistency of the target directory, and we
can even ditch support for post-fs hooks (those that restore target/).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, some filesystems may want to tweak the content of the target
directory, create special device nodes etc... This all means that:
- the content of the target directory for a specific filesystems may
depend on whether another filesystem is enabled or not; for example,
cpio will create a /init script or symlink and a /dev/console node;
- the filesystems can not be built in parallel, because they may change
the content of the target directory while another is being assembled.
Furthermore, the same fakeroot script is executed over-and-over-again
for each filesystem, to create the device nodes, the users and their
homes and files, and setting permissions...
We introduce an intermediate tarball, for which we do the full fakeroot
shebang.
That tarball then serves as the base for the other filesystems, with a
very simple fakeroot script that untars the common tarball, and calls
the actual filesystem image generator on that.
Note that we use a very simple tar command to generate the intermediate
tarball, because we are not concerned with reproducibility of the
archive itself (only of the archived files).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Before we can create an intermediate tarball for all filesystems, we
nedd to move the common dependencies needed to generate that
intermediate tarball, rather than leave those dependencies to each
filesystem.
So, we introduce rootfs-common, which gathers all those common
dependencies.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
... which for now still points to the base target directory, but this is
a step forward.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
... and locate that script in a per-rootfs directory.
Just like for ROOTFS, this variable will leak down the dependency tree to
target-finalize and packages - But it doesn't matter as it isn't used
outside fs/.
[Peter: extend commit message]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This will serve in future commits to store pre-rootfs files, like
fakeroot script...
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Like we do for packages with the PKG variable, set ROOTFS to contain the
upper-case name of the rootfs currently being generated.
This will be useful in later patches, when we need more per-rootfs
variables, like a per-rootfs TARGET_DIR for example.
In Makefiles, per-rule variables trickle down the dependency chain, to
all dependencies of that rule, so we have to stop ROOTFS as soon as
we're not in a rootfs. This means we have to stop it at target-finalize
(which is a dependency of all filesystems), and for each package
individually, since some packages (host or target) can be direct
dependencies of filesystems as well.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, the pre-gen hooks of the various filesystems are run before
we enter the fakeroot.
However, this precludes those hooks from doing actions that require
root, like creating a pseudo-device or the likes.
So, move those pre-gen hooks under fakeroot.
This has currently no side-effect, as they are still called before
everything else in the fakeroot script, even the system-wide chown call.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Like we do for post-build and post-image scripts, pass EXTRA_ENV to
post-fakeroot script.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Similar to the other compressors. Notice that we use the -l (legacy format)
for Linux kernel initrd compatibility.
Lz4 decompression is supported by the Linux kernel since 3.11.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, to register a filesystem, one has to call:
$(eval $(call ROOTFS_TARGET,blabla))
This is very unlike the package infrastructure, where the name of the
package is automatically guessed by the infra.
It turns out that we can now do that for the filesystem infra too.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some filesystems have PRE_GEN hooks that create a directory structure
under the temporary directory.
For example, iso9660 will create a sub-directory where it stores the
kernel (in case of initramfs or initrd).
So, we must run the PRE_GEN hooks after we cleanup/create the temporary
directory.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/42835965https://gitlab.com/buildroot.org/buildroot/-/jobs/42835967
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Now that they are in their own directory and no longer pollute the build
dir, there is no point in removing them.
Furthermore, a follow-up patch will require that those files survive
when more than one filesystem image is generated.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, we create a bunch of temporary files in $(BUILD_DIR), while
assembling the filesystem images.
Move those files to their own sub-directory.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Remove it just before generating the filesystem image.
This way, removing-and-recreating the file encloses the actual
image generation as tightly as possible.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The only users of post-target rules were ext2, cpio and initramfs.
Of those, ext2 and cpio were changed to use post-gen hooks, while
initramfs was not even using the generic rootfs infra and was fixed
to no longer reference post-target rules.
Besides, the comment in the infra was really misleading: it referenced
initramfs implying it was the sole user of that feature, even though
initramfs was not using the fs infra.
Furthermore, using post-target rules was inherently broken for top-level
parallel builds, because filesystems had to ensure the ordering by
themselves. Of the two real users of post-target rules (cpio and ext2),
one did enforce rules ordering (apparently correctly), while the other
forgot to do so.
We can get rid of post-target rules altogether, now.
Add a legacy check, to catch out-of-tree (e.g. br2-external) users of
post-target rules, and instruct them to switch to post-gen hooks instead.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Reviewed-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When multiple hooks are registred, both pre-a and post-command hooks'
foreach loops need to have a separator at the end in order for the
code to work as intended. Without the separator all hooks end up as a
one single line command thus making all but the first hook into
no-ops.
Fixes: 4628b6f3b4 ("fs: add pre- and post-command hooks")
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Cc: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
In some cases, the directory structure we want in the filesystem is not
exactly what we have in target/
For example, when systemd is used on a read-only rootfs, /var must be a
tmpfs. However, we may have packages that install stuff in there, and
set important rights (via the permission-table). So, at build time, we
need /var to be a symlink to the remanent location (/usr/share/factory)
while at runtime we need /var to be a directory.
One option would have been to have /var as a real directory even during
build time, and in a target-finalize hook, move everything out of there
and into the "factory" location. However, that's not possible because
it's too early: some packages may want to set ownership and/or acces
rights on directories or files in /var, and this is only done in the
fakeroot script, which is called only later during the assembling of the
filesystem images.
Also, there would have been no way to undo the tweak (i.e. we need to
restore the /var symlink so that subsequent builds continue to work) if
it were done as a target-finalize hook.
The only solution is to allow packages to register pre- and post-hooks
that are called right before and right after the rootfs commands are
executed, and inside in the fakeroot script.
We can however not re-use the BR2_ROOTFS_POST_FAKEROOT_SCRIPT feature
either because it is done before the filesystem command, but there is
nothing that is done after. Also, we don't want to add to, and modify a
user-supplied variable.
So, we introduce two new variables that packages can set to add the
commands they need to run to tweak the filesystem right at the last
moment.
Those hooks are not documented on-purpose; they are probably going to
only ever be used by systemd.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Romain Naour <romain.naour@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Since things are no longer installed in $(HOST_DIR)/usr, the callers
should also not refer to it.
This is a mechanical change with
git grep -l '$(HOST_DIR)/usr/bin' | xargs sed -i 's%$(HOST_DIR)/usr/bin%$(HOST_DIR)/bin%g'
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, when there is no syztem device table (permissions or static
devices) defined, then package permissions are not applied, because they
are guarded by the check on the system device tables being non empty.
Fix that by narrowing the guarding condition.
Note that the dependency on host-makedevs was not conditional; we always
build it even if we don't need it. Making it conditional is not
possible, because we don't know all the packages permissions by the time
the fs infra is parsed (packages from br2-external are parsed after it).
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
