]> git.ipfire.org Git - thirdparty/systemd.git/log
thirdparty/systemd.git
21 months agoci: fix a couple of typos 24671/head
Frantisek Sumsal [Wed, 14 Sep 2022 20:04:15 +0000 (22:04 +0200)] 
ci: fix a couple of typos

21 months agoci: enable a couple more possibly useful CodeQL queries
Frantisek Sumsal [Wed, 14 Sep 2022 19:32:33 +0000 (21:32 +0200)] 
ci: enable a couple more possibly useful CodeQL queries

21 months agoci: rename codeql-analysis.yml to codeql.yml
Frantisek Sumsal [Wed, 14 Sep 2022 17:13:10 +0000 (19:13 +0200)] 
ci: rename codeql-analysis.yml to codeql.yml

Just to be consistent with other repos under the systemd umbrella.

21 months agoMerge pull request #24663 from mrc0mmand/codeql-follow-up
Yu Watanabe [Tue, 13 Sep 2022 20:31:10 +0000 (05:31 +0900)] 
Merge pull request #24663 from mrc0mmand/codeql-follow-up

A couple of CodeQL tweaks and follow ups

21 months agopstore: do not try to load all known pstore modules
Nick Rosbrook [Wed, 7 Sep 2022 17:25:13 +0000 (13:25 -0400)] 
pstore: do not try to load all known pstore modules

Commit 70e74a5997 ("pstore: Run after modules are loaded") added After=
and Wants= entries for all known kernel modules providing a pstore.

While adding these dependencies on systems where one of the modules is
not present, or not configured, should not have a real affect on the
system, it can produce annoying error messages in the kernel log. E.g.
"mtd device must be supplied (device name is empty)" when the mtdpstore
module is not configured correctly.

Since dependencies cannot be removed with drop-ins, if a distro wants to
remove some of these modules from systemd-pstore.service, they need to
patch units/systemd-pstore.service.in. On the other hand, if they want
to append to the dependencies this can be done by shipping a drop-in.

Since the original intent of the previous commit was to fix [1], which
only requires the efi_pstore module, remove all other kernel module
dependencies from systemd-pstore.service, and let distros ship drop-ins
to add dependencies if needed.

[1] https://github.com/systemd/systemd/issues/18540

21 months agoMerge pull request #24662 from mrc0mmand/test-exec-deserialization-tweaks
Yu Watanabe [Tue, 13 Sep 2022 20:29:00 +0000 (05:29 +0900)] 
Merge pull request #24662 from mrc0mmand/test-exec-deserialization-tweaks

test: drop the use of `tempfile.mktemp()`

21 months agoci: limit scope for the CodeQL scan 24663/head
Frantisek Sumsal [Tue, 13 Sep 2022 19:30:10 +0000 (21:30 +0200)] 
ci: limit scope for the CodeQL scan

Don't run the workflow unnecessarily for non-{cpp,python} related changes.

21 months agoci: drop LGTM stuff and move remaining bits into a new location
Frantisek Sumsal [Tue, 13 Sep 2022 19:28:00 +0000 (21:28 +0200)] 
ci: drop LGTM stuff and move remaining bits into a new location

21 months agoci: run CodeQL on push to main/stable branches as well
Frantisek Sumsal [Tue, 13 Sep 2022 19:18:44 +0000 (21:18 +0200)] 
ci: run CodeQL on push to main/stable branches as well

Since we need results for the base branches as well in order to have
something to compare against.

Follow-up to cbe25d0dccdd3f2901a1e74a665c068f42dae9f5.

21 months agotest: drop the use of `tempfile.mktemp()` 24662/head
Frantisek Sumsal [Tue, 13 Sep 2022 18:14:48 +0000 (20:14 +0200)] 
test: drop the use of `tempfile.mktemp()`

and use `uuid.uuid4()` instead to generate a sufficiently pseudo-random
file name.

Resolves: https://github.com/systemd/systemd/security/code-scanning/142

21 months agotest: drop forgotten format()
Frantisek Sumsal [Tue, 13 Sep 2022 18:07:22 +0000 (20:07 +0200)] 
test: drop forgotten format()

Follow-up to fda00958bb08f2920cf8d42c5212fb45bdb42d6d.

21 months agoci: run CodeQL on every PR
Frantisek Sumsal [Tue, 13 Sep 2022 17:11:25 +0000 (19:11 +0200)] 
ci: run CodeQL on every PR

Since LGTM is no longer enabled for the systemd repo (as it's going to
be discontinued by the EOY), let's run CodeQL on every PR instead to
replace it.

21 months agoxdg-autostart-service: Use common boolean parser
David Edmundson [Tue, 13 Sep 2022 12:06:09 +0000 (13:06 +0100)] 
xdg-autostart-service: Use common boolean parser

Technically the desktop entry specification says value should be the
string "true" or "false". Pragmatically every desktop has their own
parsing rules which are typically less strict on how to interpret other
values.

This caused some regressions downstream when we switched to the
xdg-autostart-generator where existing handmade files contained values
with "True" or "False".

21 months agoTODO: various things about partitioning
Zbigniew Jędrzejewski-Szmek [Sun, 4 Sep 2022 08:47:28 +0000 (10:47 +0200)] 
TODO: various things about partitioning

21 months agologind: schedule idle check full interval from now if we couldn't figure out atime...
Michal Sekletar [Fri, 9 Sep 2022 11:38:58 +0000 (13:38 +0200)] 
logind: schedule idle check full interval from now if we couldn't figure out atime timestamp

21 months agoMerge pull request #24272 from dtardon/asserts
Yu Watanabe [Tue, 13 Sep 2022 12:08:10 +0000 (21:08 +0900)] 
Merge pull request #24272 from dtardon/asserts

Use ASSERT_PTR more

21 months agoboot: fix missing initialization
Yu Watanabe [Tue, 13 Sep 2022 07:57:11 +0000 (16:57 +0900)] 
boot: fix missing initialization

Fixes CID#1497847.

21 months agoMerge pull request #24651 from yuwata/openssl-util
Yu Watanabe [Tue, 13 Sep 2022 12:05:42 +0000 (21:05 +0900)] 
Merge pull request #24651 from yuwata/openssl-util

openssl-util: trivial cleanups

21 months agocreds-util: fix NULL pointer dereference
Yu Watanabe [Tue, 13 Sep 2022 08:07:50 +0000 (17:07 +0900)] 
creds-util: fix NULL pointer dereference

Fixes CID#1497840.

21 months agoopenssl-util: use assert() if no side effect 24651/head
Yu Watanabe [Tue, 13 Sep 2022 08:03:01 +0000 (17:03 +0900)] 
openssl-util: use assert() if no side effect

21 months agoopenssl-util: drop meaningless assertion
Yu Watanabe [Tue, 13 Sep 2022 08:01:26 +0000 (17:01 +0900)] 
openssl-util: drop meaningless assertion

As the type of 'msz' is int.

Fixes CID#1497842.

21 months agodbus-execute: inline variable definition 24272/head
David Tardon [Fri, 9 Sep 2022 08:31:24 +0000 (10:31 +0200)] 
dbus-execute: inline variable definition

Fixes a compiler warning:

../src/core/dbus-execute.c:460:22: error: unused variable 'c' [-Werror,-Wunused-variable]
        ExecContext *c = ASSERT_PTR(userdata);

21 months agotree-wide: check parameter before dereferencing
David Tardon [Wed, 10 Aug 2022 12:00:53 +0000 (14:00 +0200)] 
tree-wide: check parameter before dereferencing

21 months agotree-wide: use ASSERT_PTR more
David Tardon [Tue, 9 Aug 2022 12:35:15 +0000 (14:35 +0200)] 
tree-wide: use ASSERT_PTR more

21 months agoMerge pull request #23087 from yuwata/udev-watch
Yu Watanabe [Mon, 12 Sep 2022 14:41:00 +0000 (23:41 +0900)] 
Merge pull request #23087 from yuwata/udev-watch

udev: resolve race in saving inotify watch handle

21 months agoMerge pull request #23043 from yuwata/udev-node-use-flock
Yu Watanabe [Mon, 12 Sep 2022 14:40:34 +0000 (23:40 +0900)] 
Merge pull request #23043 from yuwata/udev-node-use-flock

udev-node: use flock() for symlink stack directory

21 months agobase-filesystem.c: add trailing zero byte for s390x entry
Ansgar Burchardt [Mon, 12 Sep 2022 12:55:14 +0000 (14:55 +0200)] 
base-filesystem.c: add trailing zero byte for s390x entry

21 months agohwdb: Fix Acer Aspire One AOD270/Packard Bell Dot keymappings
Hans de Goede [Mon, 29 Aug 2022 15:28:37 +0000 (17:28 +0200)] 
hwdb: Fix Acer Aspire One AOD270/Packard Bell Dot keymappings

The Acer Aspire One AOD270 and the same hardware rebranded as
Packard Bell Dot SC need a couple of keymap fixups:

1. The switch-video-mode key does not do anything. Standard acer-wmi
   maps scancode 0x61 to KEY_IGNORE since typically these events are
   duplicate with the ACPI video bus. But on these models the ACPI video
   bus does not send events for this key, so map it.

2. The Brightness up / down hotkeys send atkbd scancode 0xce / 0xef
   which by default are mapped to KEY_KPPLUSMINUS and KEY_MACRO.
   These actually are duplicate events with the ACPI video bus,
   so map these to KEY_IGNORE.

21 months agoupdate TODO
Lennart Poettering [Sun, 11 Sep 2022 23:13:44 +0000 (00:13 +0100)] 
update TODO

21 months agoupdate TODO
Lennart Poettering [Sun, 11 Sep 2022 23:13:44 +0000 (00:13 +0100)] 
update TODO

21 months agotest: add testcase for udev-watch 23087/head
Yu Watanabe [Thu, 28 Apr 2022 10:28:11 +0000 (19:28 +0900)] 
test: add testcase for udev-watch

21 months agosd-device: move device_new_from_watch_handle_at() to udev-watch.c
Yu Watanabe [Thu, 14 Apr 2022 21:38:33 +0000 (06:38 +0900)] 
sd-device: move device_new_from_watch_handle_at() to udev-watch.c

And drop unused watch handle related functions.

21 months agoudev: warn on udev_watch_{begin,end}() failure
Yu Watanabe [Thu, 28 Apr 2022 07:42:12 +0000 (16:42 +0900)] 
udev: warn on udev_watch_{begin,end}() failure

21 months agoudev: drop unnecessary call of udev_watch_end()
Yu Watanabe [Thu, 28 Apr 2022 07:39:49 +0000 (16:39 +0900)] 
udev: drop unnecessary call of udev_watch_end()

As it is already called by udev_event_execute_rules().

21 months agoudev: use rm_rf() to remove old watch directory
Yu Watanabe [Thu, 28 Apr 2022 06:54:06 +0000 (15:54 +0900)] 
udev: use rm_rf() to remove old watch directory

21 months agoudev-watch: remove symlink for saving inotify watch handle only when it is owned...
Yu Watanabe [Thu, 14 Apr 2022 21:31:21 +0000 (06:31 +0900)] 
udev-watch: remove symlink for saving inotify watch handle only when it is owned by the processing device

Before removing symlinks that stores watch handles, this makes udev
worker check if the symlink is owned by the processing device.
Then, we can avoid TOCTOU and drop the try-and-wait loop.

This partially reverts 2d3af41f0e837390b734253f5c4a99a9f33c53e3.

21 months agoudev: ignore IN_IGNORED inotify event earlier
Yu Watanabe [Fri, 15 Apr 2022 00:42:15 +0000 (09:42 +0900)] 
udev: ignore IN_IGNORED inotify event earlier

21 months agoudev: not necessary to return 1 from on_inotify()
Yu Watanabe [Thu, 14 Apr 2022 21:08:13 +0000 (06:08 +0900)] 
udev: not necessary to return 1 from on_inotify()

21 months agoMerge pull request #24637 from mrc0mmand/TEST-75-tweaks
Yu Watanabe [Sun, 11 Sep 2022 16:17:36 +0000 (01:17 +0900)] 
Merge pull request #24637 from mrc0mmand/TEST-75-tweaks

test: make the resolved notifications check a bit more robust

21 months agotest: make the resolved notifications check a bit more robust 24637/head
Frantisek Sumsal [Sun, 11 Sep 2022 12:17:56 +0000 (14:17 +0200)] 
test: make the resolved notifications check a bit more robust

Let's parse the resolved JSON notifications via `jq` and check them in a
bit more "controlled" manner - e.g. until now the `grep` was checking just
a one gigantic JSON string, as all received notifications via the
varlink socket are terminated by a NUL character, not a newline.

Also, as the notification delivery is asynchronous, retry the check
a couple of times if it fails (spotted in C8S jobs):

```
[ 2891.935879] testsuite-75.sh[36]: + : '--- nss-resolve/nss-myhostname tests'
[ 2891.935988] testsuite-75.sh[36]: + run getent -s resolve hosts ns1.unsigned.test
[ 2891.936542] testsuite-75.sh[177]: + getent -s resolve hosts ns1.unsigned.test
[ 2891.937499] testsuite-75.sh[178]: + tee /tmp/tmp.pqjNvbQ2eS
[ 2891.939977] testsuite-75.sh[178]: 10.0.0.1        ns1.unsigned.test
[ 2891.940258] testsuite-75.sh[36]: + grep -qE '^10\.0\.0\.1\s+ns1\.unsigned\.test' /tmp/tmp.pqjNvbQ2eS
[ 2891.942235] testsuite-75.sh[189]: + grep -qF '[10,0,0,1]'
[ 2891.942577] testsuite-75.sh[188]: + grep -aF ns1.unsigned.test /tmp/notifications.txt
[ 2891.943978] systemd[1]: testsuite-75.service: Child 36 belongs to testsuite-75.service.
[ 2891.944112] systemd[1]: testsuite-75.service: Main process exited, code=exited, status=1/FAILURE
[ 2891.944215] systemd[1]: testsuite-75.service: Failed with result 'exit-code'.
```

21 months agotest: install jq into all images
Frantisek Sumsal [Sun, 11 Sep 2022 12:16:34 +0000 (14:16 +0200)] 
test: install jq into all images

21 months agoMerge pull request #24593 from jamacku/main
Frantisek Sumsal [Sat, 10 Sep 2022 19:42:26 +0000 (19:42 +0000)] 
Merge pull request #24593 from jamacku/main

Update `advanced-issue-labeler` action - `@v1` -> `@v2`

21 months agotpm2-util: avoid trial
Lennart Poettering [Fri, 9 Sep 2022 21:06:58 +0000 (23:06 +0200)] 
tpm2-util: avoid trial

Let's avoid trial mode for putting together policies if we can, and use
the real policy whenever we can pull it off.

Fixes: #24401
21 months agotest-29-portable: set timeout for 'portablectl reattach'
Yu Watanabe [Fri, 9 Sep 2022 18:21:37 +0000 (03:21 +0900)] 
test-29-portable: set timeout for 'portablectl reattach'

Then, the test can fail earlier than the timeout of the whole test
specified by $QEMU_TIMEOUT=.

This is useful when we try to run the test multiple times.

Workaround for issue #24147.

21 months agoMerge pull request #24618 from yuwata/udev-split-synthesizing
Yu Watanabe [Fri, 9 Sep 2022 17:52:54 +0000 (02:52 +0900)] 
Merge pull request #24618 from yuwata/udev-split-synthesizing

udev: shorten synthesize_change() a bit

21 months agoblockdev-util: split out blockdev_reread_partition_table() 24618/head
Yu Watanabe [Fri, 9 Sep 2022 15:33:43 +0000 (00:33 +0900)] 
blockdev-util: split out blockdev_reread_partition_table()

No functional changes, just refactoring.

21 months agoMerge pull request #24458 from poettering/stub-embedd-sig
Lennart Poettering [Fri, 9 Sep 2022 16:18:37 +0000 (18:18 +0200)] 
Merge pull request #24458 from poettering/stub-embedd-sig

optionally, embed PCR signature and public key in new sd-stub PE sections

21 months agoudevd: use partition enumerator at one more place
Yu Watanabe [Fri, 9 Sep 2022 15:17:46 +0000 (00:17 +0900)] 
udevd: use partition enumerator at one more place

No functional changes, just refactoring.

21 months agoMerge pull request #24609 from yuwata/sd-device-enumerator-check-sysname
Yu Watanabe [Fri, 9 Sep 2022 14:32:15 +0000 (23:32 +0900)] 
Merge pull request #24609 from yuwata/sd-device-enumerator-check-sysname

sd-device-enumerator: check sysname earlier

21 months agoMerge pull request #24611 from yuwata/loop-util-refactoring
Yu Watanabe [Fri, 9 Sep 2022 14:31:19 +0000 (23:31 +0900)] 
Merge pull request #24611 from yuwata/loop-util-refactoring

loop-util: several cleanups

21 months agoblockdevi-util: also check sysname when enumerating partitions
Yu Watanabe [Thu, 8 Sep 2022 20:25:55 +0000 (05:25 +0900)] 
blockdevi-util: also check sysname when enumerating partitions

21 months agosd-device-enumerator: use _cleanup_free_ attribute for safety 24609/head
Yu Watanabe [Fri, 9 Sep 2022 11:55:45 +0000 (20:55 +0900)] 
sd-device-enumerator: use _cleanup_free_ attribute for safety

No functional changes, just improve readability.

Suggested by https://github.com/systemd/systemd/pull/24601#discussion_r966883459.

21 months agosd-device-enumerator: always ignore ENOENT from opendir()
Yu Watanabe [Fri, 9 Sep 2022 11:52:41 +0000 (20:52 +0900)] 
sd-device-enumerator: always ignore ENOENT from opendir()

But logs all errors.

21 months agosd-device-enumerator: check sysname earlier
Yu Watanabe [Thu, 8 Sep 2022 19:26:15 +0000 (04:26 +0900)] 
sd-device-enumerator: check sysname earlier

21 months agoupdate TODO 24458/head
Lennart Poettering [Fri, 26 Aug 2022 09:11:07 +0000 (11:11 +0200)] 
update TODO

21 months agoman: say early what a unified kernel image/UKI is
Lennart Poettering [Fri, 9 Sep 2022 09:24:12 +0000 (11:24 +0200)] 
man: say early what a unified kernel image/UKI is

21 months agoman: be more careful regarding TPM vs. TPM2
Lennart Poettering [Fri, 9 Sep 2022 09:09:30 +0000 (11:09 +0200)] 
man: be more careful regarding TPM vs. TPM2

We support PCR measurements for both classic TPM1.2 and TPM2, hence just
say "TPM" generically in that context. But the signed policies are
exclusive to TPM2, hence always say TPM2 there.

We mostly got that right, except at one place. Fix that.

21 months agoman: document the new .pcrsig/.pcrpkey sections for unified kernel images
Lennart Poettering [Fri, 9 Sep 2022 09:08:35 +0000 (11:08 +0200)] 
man: document the new .pcrsig/.pcrpkey sections for unified kernel images

21 months agotmpfiles: copy PCR sig/pkey from initrd /.extra/ into /run/
Lennart Poettering [Thu, 25 Aug 2022 15:16:52 +0000 (17:16 +0200)] 
tmpfiles: copy PCR sig/pkey from initrd /.extra/ into /run/

Now that sd-stub will place the PCR signature and its public key in
the initrd's /.extra/ directory, let's copy it from there into /run/
from userspace. This is done because /.extra/ is on the initrd's tmpfs
which will be emptied during the initrd → host transition. Since we want
these two files to survive we'll copy them – if they exist – into /run/
where they will survive the transition.

Thus, with this last change the files will have safely propagated from
their PE sections into files in /run/ where userspace can find them

The paths in /run/ happen to be the exact ones that
systemd-cryptenroll/systemd-cryptsetup/systemd-creds look for them.

21 months agostub: pass .pcrsig and .pcrpkey PE sections as cpio into invoked kernel
Lennart Poettering [Thu, 25 Aug 2022 14:55:49 +0000 (16:55 +0200)] 
stub: pass .pcrsig and .pcrpkey PE sections as cpio into invoked kernel

Pick up the two new sections in sd-stub and pass them as initrds into
the booted kernels, where they'll show up as
/.extra/tpm2-pcr-signature.json and /.extra/tpm2-pcr-public-key.pem in
the initrd file system.

The initrd is then supposed to pick these files up from there and save
them at a place that will survive into the host OS.

21 months agostub: add new special PE sections ".pcrsig" and ".pcrpkey" in unified kernels
Lennart Poettering [Thu, 25 Aug 2022 14:55:01 +0000 (16:55 +0200)] 
stub: add new special PE sections ".pcrsig" and ".pcrpkey" in unified kernels

These aren't wired up yet to do anything useful. For now we just define
them.

This sections are supposed to carry a signature for expected
measurements on PCR 11 if this kernel is booted, in the JSON format
"systemd-measure sign" generates, and the public key used for the
signature.

The idea is to embedd the signature and the public key in unified
kernels and making them available to userspace, so that userspace can
easily access them and enroll (for which the public key is needed) or
unlock (for which the PCR signature is needed) LUKS2 volumes and
credentials that are bound to the currently used kernel version stream.

Why put these files in PE sections rather than just into simple files in
the initrd or into the host fs?

The signature cannot be in the initrd, since it is after all covering
the initrd, and thus the initrd as input for the calculation cannot
carry the result of the calculation. Putting the signature onto the root
fs sucks too, since we typically want to unlock the root fs with it,
hence it would be inaccessible for it's primary purpose then.

The public key could be in the initrd or in the root fs, there's no
technical restriction for that. However, I still think it's a good idea
to put it in a PE section as well, because this means the piece of code
that attaches the signature can also attach the public key easily in one
step, which is nice since it allows separating the roles of the
kernel/initrd/root fs builder, and the role of the signer, and the
former doesn't have to have knowledge about what the latter is going to
add to the image.

Note that the signature section is excluded from the TPM measurements
sd-stub does about its resource sections, since – as mentioned – it's
the expected output of the signing operation whose input are the
measurements, hence it cannot also be input to them. The public key
section is included in the measurements however.

21 months agocpio: add helper for packing cpios of literally specified data blobs
Lennart Poettering [Thu, 25 Aug 2022 14:54:03 +0000 (16:54 +0200)] 
cpio: add helper for packing cpios of literally specified data blobs

Let's add simple helpers for passing data blobs from the stub into the
booted kernel as initrds that are generated on-the-fly.

(Note used yet, a later commit will make use of this)

21 months agoImplement DNS notifications from resolved via varlink
Suraj Krishnan [Tue, 26 Apr 2022 22:09:02 +0000 (17:09 -0500)] 
Implement DNS notifications from resolved via varlink

* The new varlink interface exposes a method to subscribe to DNS
resolutions on the system. The socket permissions are open for owner and
group only.
* Notifications are sent to subscriber(s), if any, after successful
resolution of A and AAAA records.

This feature could be used by applications for auditing/logging services
downstream of the resolver. It could also be used to asynchronously
update the firewall. For example, a system that has a tightly configured
firewall could open up connections selectively to known good hosts based
on a known allow-list of hostnames. Of course, updating the firewall
asynchronously will require other design considerations (such as
queueing packets in the user space while a verdict is made).

See also:
https://lists.freedesktop.org/archives/systemd-devel/2022-August/048202.html
https://lists.freedesktop.org/archives/systemd-devel/2022-February/047441.html

21 months agoMerge pull request #24557 from yuwata/repart
Yu Watanabe [Fri, 9 Sep 2022 00:32:07 +0000 (09:32 +0900)] 
Merge pull request #24557 from yuwata/repart

repart: fix partition size calculation

21 months agotest-58-repart: add test case for issue #24553 24557/head
Yu Watanabe [Mon, 5 Sep 2022 13:11:55 +0000 (22:11 +0900)] 
test-58-repart: add test case for issue #24553

21 months agorepart: make existing partition can be also 'dropped'
Yu Watanabe [Mon, 5 Sep 2022 16:58:34 +0000 (01:58 +0900)] 
repart: make existing partition can be also 'dropped'

Previously, when an existing partition cannot grow, then entire process
fails. This makes such an existing partion handled as an foreign
partition, i.e. it is not managed by us.

21 months agorepart: check if existing partitions can grow
Yu Watanabe [Mon, 5 Sep 2022 12:57:38 +0000 (21:57 +0900)] 
repart: check if existing partitions can grow

Fixes #24553.

21 months agorepart: split out free_area_{current,min}_end() from free_area_available_for_new_part...
Yu Watanabe [Mon, 5 Sep 2022 12:56:49 +0000 (21:56 +0900)] 
repart: split out free_area_{current,min}_end() from free_area_available_for_new_partitions()

No actual code changes, just preparation for later commits.

21 months agorepart: reset assignments by previous context_allocate_partitions()
Yu Watanabe [Mon, 5 Sep 2022 16:23:19 +0000 (01:23 +0900)] 
repart: reset assignments by previous context_allocate_partitions()

The function context_allocate_partitions() may be called multiple times.
If this is called multiple times, then dropped partitions may still
assigned to free area.

21 months agorepart: do not assign new size larger than acquired or the specified maximum
Yu Watanabe [Sun, 4 Sep 2022 09:54:52 +0000 (18:54 +0900)] 
repart: do not assign new size larger than acquired or the specified maximum

The acquired size may be larger than the requested maximum. So, let's
cap the value.

Note, at the final phase, the acquired size should be larger than the
requested minimum. Hence, the assertion about that is added.

21 months agorepart: set new size for foreign partitions at first
Yu Watanabe [Sun, 4 Sep 2022 09:48:15 +0000 (18:48 +0900)] 
repart: set new size for foreign partitions at first

Otherwise, the new size may be larger than the acquired one.

21 months agorepart: anyway run loop at the end even if the loop will be restarted later
Yu Watanabe [Sun, 4 Sep 2022 09:51:55 +0000 (18:51 +0900)] 
repart: anyway run loop at the end even if the loop will be restarted later

The order of the partitions processed in each phase does not change
result for the first two phase (PHASE_OVERCHARGE and PHASE_UNDERCHARGE).

21 months agorepart: make scale_by_weight() always succeed
Yu Watanabe [Sun, 4 Sep 2022 06:20:29 +0000 (15:20 +0900)] 
repart: make scale_by_weight() always succeed

21 months agorepart: split out context_grow_partition_one()
Yu Watanabe [Sun, 4 Sep 2022 09:43:55 +0000 (18:43 +0900)] 
repart: split out context_grow_partition_one()

No actual code changes, just refactoring.

21 months agorepart: introduce partition_{min,max}_padding()
Yu Watanabe [Sun, 4 Sep 2022 09:26:04 +0000 (18:26 +0900)] 
repart: introduce partition_{min,max}_padding()

No actual code changes, just refactoring.

21 months agorepart: ensure partition_max_size() >= partition_min_size()
Yu Watanabe [Sun, 4 Sep 2022 09:25:03 +0000 (18:25 +0900)] 
repart: ensure partition_max_size() >= partition_min_size()

21 months agorepart: make partition_max_size() return UINT64_MAX if not specified
Yu Watanabe [Sun, 4 Sep 2022 06:19:18 +0000 (15:19 +0900)] 
repart: make partition_max_size() return UINT64_MAX if not specified

Previously, it did not return UINT64_MAX, but a huge value, as
`UINT64_MAX / grain_size * grain_size != UINT64_MAX`.

This also drops unnecessary conditions.

21 months agorepart: constify partition_min_size()
Yu Watanabe [Sun, 4 Sep 2022 09:24:13 +0000 (18:24 +0900)] 
repart: constify partition_min_size()

21 months agorepart: rename variables in config_parse_weight()
Yu Watanabe [Mon, 5 Sep 2022 16:42:44 +0000 (01:42 +0900)] 
repart: rename variables in config_parse_weight()

This is for Weight= or PaddingWeight=, not for Priority=.

No actual code changes, just refactoring.

21 months agoloop-util: do not try to enumerate partitions twice 24611/head
Yu Watanabe [Thu, 8 Sep 2022 20:12:07 +0000 (05:12 +0900)] 
loop-util: do not try to enumerate partitions twice

21 months agoloop-util: always check if backing file is not attached yet
Yu Watanabe [Thu, 8 Sep 2022 20:04:08 +0000 (05:04 +0900)] 
loop-util: always check if backing file is not attached yet

An image file without partition table may be attached.

21 months agoloop-util: open loopback block device in loop_configure()
Yu Watanabe [Thu, 8 Sep 2022 08:18:49 +0000 (17:18 +0900)] 
loop-util: open loopback block device in loop_configure()

And make it return LoopDevice object on success.
No functional changes, just refactoring.

21 months agoloop-util: save the flag about if LOOP_CONFIGURE ioctl works or not in loop_configure()
Yu Watanabe [Thu, 8 Sep 2022 08:01:45 +0000 (17:01 +0900)] 
loop-util: save the flag about if LOOP_CONFIGURE ioctl works or not in loop_configure()

When LOOP_CONFIGURE fail, then it is not necessary to call it on any
later call of loop_configure(). Let's save the flag in a static variable
in the function.

21 months agoloop-util: split out several functions fron loop_configure() and loop_device_make_int...
Yu Watanabe [Thu, 8 Sep 2022 07:44:40 +0000 (16:44 +0900)] 
loop-util: split out several functions fron loop_configure() and loop_device_make_internal()

The two functions were quite long, and hard to understand its logic.
No functional change, just refactoring.

21 months agoMerge pull request #24601 from yuwata/sd-device-enumerator-drop-recursion
Yu Watanabe [Thu, 8 Sep 2022 19:30:42 +0000 (04:30 +0900)] 
Merge pull request #24601 from yuwata/sd-device-enumerator-drop-recursion

sd-device-enumerator: do not recursively find child devices

21 months agoMerge pull request #24351 from poettering/pcr-sign
Frantisek Sumsal [Thu, 8 Sep 2022 19:07:04 +0000 (19:07 +0000)] 
Merge pull request #24351 from poettering/pcr-sign

support for signed TPM2 PCR policies in cryptsetup/cryptenrolls/credentials

21 months agomeasure: print a helpful message if TPM2 PCR sysfs attributes are missing
Lennart Poettering [Thu, 8 Sep 2022 10:57:23 +0000 (12:57 +0200)] 
measure: print a helpful message if TPM2 PCR sysfs attributes are missing

Prompted by:

https://github.com/systemd/systemd/pull/24351#pullrequestreview-1100418585

21 months agobasic/missing_loop.h: fix missing lo_flags LO_FLAGS_DIRECT_IO
Guillaume W. Bres [Thu, 8 Sep 2022 13:54:57 +0000 (15:54 +0200)] 
basic/missing_loop.h: fix missing lo_flags LO_FLAGS_DIRECT_IO

21 months agoMerge pull request #24608 from bluca/example
Luca Boccassi [Thu, 8 Sep 2022 16:42:07 +0000 (17:42 +0100)] 
Merge pull request #24608 from bluca/example

docs: mention tmpfiles.d in CREDENTIALS.md and add example for ssh provisioning (to manpage too)

21 months agoMerge pull request #24600 from poettering/bootspec-chase-symlinks
Lennart Poettering [Thu, 8 Sep 2022 16:40:10 +0000 (18:40 +0200)] 
Merge pull request #24600 from poettering/bootspec-chase-symlinks

bootspec: more chase_symlinks()

21 months agodocs: mention tmpfiles.d in CREDENTIALS.md and add example 24608/head
Luca Boccassi [Thu, 8 Sep 2022 15:33:24 +0000 (16:33 +0100)] 
docs: mention tmpfiles.d in CREDENTIALS.md and add example

21 months agoman: add example with one-liner for ssh provisioning via tmpfiles.d + Creds
Luca Boccassi [Thu, 8 Sep 2022 15:27:52 +0000 (16:27 +0100)] 
man: add example with one-liner for ssh provisioning via tmpfiles.d + Creds

21 months agotest: check the numa-test.service status directly
Frantisek Sumsal [Thu, 8 Sep 2022 10:27:46 +0000 (12:27 +0200)] 
test: check the numa-test.service status directly

In a couple of recent CI runs I noticed TEST-36 failing due to a missed
service exit notification and a subsequent fail of the `grep` command:

```
[  257.112153] H systemd[1]: Started numa-test.service.
[  257.114343] H systemd[899]: numa-test.service: Failed to set NUMA memory policy: Invalid argument
[  257.118270] H systemd[899]: numa-test.service: Failed at step NUMA_POLICY spawning /bin/sleep: Invalid argument
[  257.126170] H systemd[1]: Bus private-bus-connection: changing state RUNNING → CLOSING
[  257.130290] H systemd[1]: numa-test.service: Failed to send unit change signal for numa-test.service: Connection reset by peer
[  257.131567] H systemd[1]: Received SIGCHLD from PID 899 ((sleep)).
[  257.132870] H systemd[1]: Got disconnect on private connection.
[  257.134299] H systemd[1]: systemd-journald.service: Got notification message from PID 498 (FDSTORE=1)
[  257.135611] H systemd[1]: systemd-journald.service: Added fd 38 (n/a) to fd store.
[  257.136999] H systemd[1]: systemd-journald.service: Received EPOLLHUP on stored fd 38 (stored), closing.
[  257.366996] H testsuite-36.sh[536]: + stopJournalctl
[  257.366996] H testsuite-36.sh[536]: + local unit=init.scope
[  257.366996] H testsuite-36.sh[536]: + echo 'Force journald to write all queued messages'
[  257.366996] H testsuite-36.sh[536]: Force journald to write all queued messages
[  257.366996] H testsuite-36.sh[536]: + journalctl --sync
[  257.488642] H systemd-journald[498]: Received client request to rotate journal.
[  257.520821] H testsuite-36.sh[536]: + journalctl -u init.scope --cursor-file=jounalCursorFile
[  257.981399] H testsuite-36.sh[536]: + pid1StopUnit numa-test.service
[  257.984533] H testsuite-36.sh[536]: + systemctl stop numa-test.service
[  258.173656] H systemd[1]: Bus private-bus-connection: changing state AUTHENTICATING → RUNNING
[  258.180710] H systemd[1]: numa-test.service: Trying to enqueue job numa-test.service/stop/replace
[  258.182424] H systemd[1]: Added job numa-test.service/stop to transaction.
[  258.185234] H systemd[1]: numa-test.service: Installed new job numa-test.service/stop as 738
[  258.187017] H systemd[1]: numa-test.service: Enqueued job numa-test.service/stop as 738
[  258.239930] H testsuite-36.sh[536]: + grep 'numa-test.service: Main process exited, code=exited, status=242/NUMA' journal.log
```

Let's mitigate this by checking the test service exit status directly
instead of relying on the notification.

21 months agoTEST-70-TPM2: add test for new signed TPM2 PCR policies 24351/head
Lennart Poettering [Fri, 19 Aug 2022 09:26:49 +0000 (11:26 +0200)] 
TEST-70-TPM2: add test for new signed TPM2 PCR policies

21 months agocryptsetup: hook up TPM2 token code with policies based on PCR signatures, too
Lennart Poettering [Fri, 19 Aug 2022 20:18:40 +0000 (22:18 +0200)] 
cryptsetup: hook up TPM2 token code with policies based on PCR signatures, too

21 months agotpm2-util: add helper for formatting PCR masks as string
Lennart Poettering [Fri, 19 Aug 2022 20:18:31 +0000 (22:18 +0200)] 
tpm2-util: add helper for formatting PCR masks as string

21 months agotpm2-util: add common parser for the LUKS2 TPM2 JSON structure
Lennart Poettering [Fri, 19 Aug 2022 20:15:12 +0000 (22:15 +0200)] 
tpm2-util: add common parser for the LUKS2 TPM2 JSON structure

This splits out the JSON parser used by the systemd-cryptsetup code.

This is preparation for later work to reuse it in the tpm2 cryptsetup
token module, which currently uses a separate but very similar parser
for the same data.

No change in behaviour.

21 months agocryptsetup: hook up signed PCR policies
Lennart Poettering [Thu, 18 Aug 2022 09:10:30 +0000 (11:10 +0200)] 
cryptsetup: hook up signed PCR policies

21 months agorepart: hook up new TPM2 signed policies with repart
Lennart Poettering [Wed, 17 Aug 2022 15:30:11 +0000 (17:30 +0200)] 
repart: hook up new TPM2 signed policies with repart

21 months agocryptenroll: hook up new TPM2 signed policies with cryptenroll
Lennart Poettering [Wed, 17 Aug 2022 15:29:44 +0000 (17:29 +0200)] 
cryptenroll: hook up new TPM2 signed policies with cryptenroll