For repo structure directories like `objects`, `refs`, etc... we should
be more permissive and let the system's `umask` narrow down the
permission bits as wanted.
This came up in a context where we want to be able to have read/write
access on an OSTree repo on NFS from two separate OpenShift apps by
using supplemental groups[1] so we don't require SCCs for running as the
same UID (supplemental groups are part of the default restricted SCC).
[1] https://docs.openshift.com/container-platform/3.11/install_config/persistent_storage/persistent_storage_nfs.html#nfs-supplemental-groups
For some reason I haven't fully debugged (probably a recent
kernel change), in the case where the immutable bit isn't set,
trying to call `EXT2_IOC_SETFLAGS` without it set returns `EINVAL`.
Let's avoid calling the `ioctl()` if we don't have anything to do.
This fixes a slew of `make check` failures here in my toolbox
environment.
(kernel is `5.5.0-0.rc6.git0.1.fc32.x86_64` with `xfs`)
Move the alternative builds into the Jenkinsfile.
Update it to do a FCOS build + kola run.
We drop the flatpak/rpm-ostree runs for now; the former
will needs some work to do the automatic virt bits. The
latter I think we can circle back to when we e.g. figure
out how to include rpm-ostree's tests in kola runs.
Using fs-verity is natural for OSTree because it's file-based,
as opposed to block based (like dm-verity). This only covers
files - not symlinks or directories. And we clearly need to
have integrity for the deployment directories at least.
Also, what we likely need is an API that supports signing files
as they're committed.
So making this truly secure would need a lot more work. Nevertheless,
I think it's time to start experimenting with it. Among other things,
it does *finally* add an API that makes files immutable, which will
help against some accidental damage.
This is basic enablement work that is being driven by
Fedora CoreOS; see also https://github.com/coreos/coreos-assembler/pull/876
17db0f15a7 ("configure: add option for libsystemd") exposed
--without-libsystemd to allow systemd to be disabled even if the systemd
pkgconfig script was present, introducing a new variable
with_libsystemd; there are now three, almost identical variables:
- with_libsystemd [yes, no, maybe] - controlled by --without-libsystemd,
resolved into yes/no by the initial checks
- have_libsystemd [yes, no, <undefined>] - only set if with_libsystemd
is yes/maybe, otherwise undefined
- with_systemd [yes, <undefined>] - yes if have_systemd is yes,
otherwise undefined
with_systemd is the earliest variable and was previously set by a set of
checks for dracut and mkinitcpio. These checks were changed for a
systemd check in 9e2763106b ("lib: Use sd_journal directly
(optionally)"). This commit also introduced BUILDOPT_LIBSYSTEMD, which
will always match BUILDOPT_SYSTEMD.
Fix the confusion by removing with_systemd which will always be yes when
with_libsystemd=yes, or undefined if with_libsystemd=no. We can ignore
the with_libsystemd=maybe case because it will always be resolved into
yes/no before with_systemd is set.
And replace all uses of BUILDOPT_SYSTEMD with BUILDOPT_LIBSYSTEMD, since
they again always match.
This fixes both the advertised features and the summary output when
systemd is disabled by using with_libsystemd which is always defined.
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Fixes: 5c62a7e4d0 ("build: Expose systemd in OSTREE_FEATURES")
Fixes: 17db0f15a7 ("configure: add option for libsystemd")
Supersedes: #1992
When --autoexit is used with --daemonize and --log-file, the program
never exits when the root directory is deleted. I believe what happens
is that g_file_new_for_path triggers the glib worker context to be
started to talk to GVfs. Once the program forks, the parent exits and
the thread iterating the worker context is gone. The file monitor then
never receives any events because the inotify helper also runs from the
worker context.
Move the fork earlier just after parsing and validating the command line
arguments. In order to handle setup errors in the child, a pipe is
opened and the parents waits until the child writes a status byte to it.
If the byte is 0, the parent considers the child setup successful and
exits and the child carries on as a daemon. Notably, the child doesn't
reopen stderr to /dev/null until after this so that it can send error
messages there.
Fixes: #1941
Some of the flatpak tests assert on GPG error strings that come from
OSTree. Those are being changed here, so patch the cloned flatpak 1.4.1
to accommodate the new error strings. When this work lands, I'll send a
patch upstream to flatpak that will eventually trickle back here in a
tagged build.
This checks whether gpg-import will properly update the keyring for a
key that already exists. In particular, we check that changing the key
expiration time or revoking it results in commit verification failure
after re-importing the keys.
Add explicit tests for
`ostree_gpg_verify_result_require_valid_signature` in addition to the
implicit tests via `ostree pull` and others. This allows checking the
error code raised.
Currently `ostree_gpg_verify_result_require_valid_signature` always
returns an error that the key used for the signature is missing from the
keyring. However, all that's been determined is that there are no valid
signatures. The error could also be from an expired signature, an
expired key, a revoked key or an invalid signature.
Provide values for these missing errors and return them from
`ostree_gpg_verify_result_require_valid_signature`. The description of
each result is appended to the error message, but since the result can
contain more than one signature but only a single error can be returned,
the status of the last signature is used for the error code. See the
comment for rationale.
Related: flatpak/flatpak#1450
Currently tests are always run against the full lgpl2.sig file with all
signatures, but it should also be possible to specify one or more of the
individual lgpgl2.sig<N> files.
Drop the current usage of passing the signature index in the test data
since it's always specific to the test function and instead provide an
optional array of signature files for the test fixture to sign with.
When the private keys were generated, gpg added an ultimate trust entry
since you normally want to trust your own keys. However, this throws off
the expired signature testing since gpgme considers it valid if the key
is fully or ultimately trusted.
The use of a trustdb for the test-gpg-verify-result is unlike any other
GPG verification in ostree. Under normal circumstances, a temporary GPG
homedir is created without any trust information, so all keys are
treated as having unknown trust.
Regenerate an empty trustdb.gpg in gpg-verify-data so that the tests
behave as ostree normally operates. After this the expired signature
testing correctly shows up as a non-valid signature. The trustdb was
regenerated by simply removing it and running any gpg operation with the
gpg-verify-data directory as the homedir.
The full block with all 5 signatures remains, but this allows passing
individual signatures through the GPG verification APIs. The split was
done with `gpgsplit`, and looking at the output of `gpg --list-packets`
of the split and unsplit files appears correct.
These can then be imported during a test to revoke a key without trying
to go through the gpg --generate-revocation dialog. Note that these need
to go in a subdirectory of the homedir since `gpgkeypath` will try to
import every regular file in the homedir.
gpg prints a warning about unsafe permissions if the homedir is group or
world readable. This is just noise in the test logs, so appease it by
making the homedir 700.
Use long GPG key IDs as it's safer and matches the format used by gpg
and gpgme. Add the associated fingerprints since these are needed by gpg
when manipulating keys.
The vmcheck tests in 2019.3 fail because of an SSH control socket issue
on overlayfs. This is fixed in 2019.4[1]. That has some other changes
such as using Python 3 in tests. The package dependencies have been
synced from the rpm-ostree CI for that.
Unfortunately, this is no longer a totally representative test of f29
since it has 2019.3 in updates. But that's the price you pay for
exercising someone else's CI from your own CI.
1. c89f81c138Fixes: #1994
The case-ignoring regex `^(C|en_US)` will match any locale that starts
with `c`. On my system this is `ca_AD.utf8`, which breaks the test
suite. Instead, use a single regex that includes the joining `.` rather
than 2 separate regexes. This also changes `head` to use the `-n`
option, which has been preferred for at least 10 years in the coreutils
version and is supported by busybox as well.
This function parses the object listing in the `ostree.sizes` metadata
and returns an array of `OstreeCommitSizesEntry` structures.
Unfortunately, for reasons I don't understand, the linker wants to
resolve `_ostree_read_varuint64` from `ostree-core.c` even though it's
not used by `test-checksum.c` at all.
Append a byte encoding the OSTree object type for each object in the
metadata. This allows the commit metadata to be fetched and then for the
program to see which objects it already has for an accurate calculation
of which objects need to be downloaded.
This slightly breaks the `ostree.sizes` `ay` metadata entries. However,
it's unlikely anyone was asserting the length of the entries since the
array currently ends in 2 variable length integers. As far as I know,
the only users of the sizes metadata are the ostree test suite and
Endless' eos-updater[1]. The former is updated here and the latter
already expects this format.
1. https://github.com/endlessm/eos-updater/
If the object was already in the repo then the sizes metadata entry was
skipped. Move the sizes entry creation after the data has been computed
but before the early return for an existing object.
The object sizes hash table was only being cleared when the repo was
finalized. That means that performing multiple commits while the repo
was open would reuse all the object sizes metadata for each commit.
Clear the hash table when the sizes metadata is setup and when it's
added to a commit. This still does not fix the issue all the way since
it does nothing to prevent the program from constructing multiple
commits simultaneously. To handle that, the object sizes hash table
should be attached to the MutableTree since that has the commit state.
However, the MutableTree is gone when the commit is actually created.
The hash table would have to be transferred to the root file when
writing the MutableTree. That would be an awkward addition to
OstreeRepoFile, though. Add a FIXME to capture that.
Ensure all 3 of the checksum, compressed size and uncompressed size are
correct. For repeatable objects, skip xattrs and use canonical
permissions for the commit. For the sizes, read a varint rather than
assuming they will be a single byte. To work around bugs in gjs with
byte array unpacking, manually build the array byte by byte. Split out
some helper functions to use in subsequent tests.
Jonathan Lebon
OpenShift Merge Robot
Colin Walters
Cole Robinson
Alex Kiernan
Dan Nicholson
John Hiesey