Benjamin Berg [Mon, 10 Feb 2020 14:53:55 +0000 (15:53 +0100)]
docs: Add some notes about managing graphical user sessions
This is work in progress and not finished yet. However, I hope to have
captured some of the key points that came up in previous discussions
with appropriate notes about things that still need to be defined.
I may revisit it later. Also, feel free to completely rewrite if the
format is not quite right.
Balint Reczey [Wed, 18 Mar 2020 17:29:02 +0000 (18:29 +0100)]
user-util: Allow names starting with a digit
In 1a29610f5fa1bcb2eeb37d2c6b79d8d1a6dbb865 the change inadvertedly
disabled names with digit as the first character. This follow-up change
allows a digit as the first character in compat mode.
Wen Yang [Wed, 25 Mar 2020 09:08:50 +0000 (17:08 +0800)]
core/unit: print info when unit_add_name failed
When there are hundreds of mounts on the server, it will take a long
time to analyze the failure of a certain mount unit. So it is useful
to print the reason why unit_add_name() failed.
Add a simple check on the number of unit files that were found: return
`-ENOENT` when none is found from the function and thus `systemctl`
consequently exits with `1` (`EXIT_FAILURE`) if none were found.
Verification:
```bash
root@image:~# systemctl list-unit-files dbus-nonexistant.service; echo
$?
UNIT FILE STATE VENDOR PRESET
0 unit files listed.
1
root@image:~# systemctl list-unit-files dbus.service; echo $?
UNIT FILE STATE VENDOR PRESET
dbus.service static enabled
Anita Zhang [Sun, 8 Mar 2020 01:14:35 +0000 (17:14 -0800)]
cgroup-util: helper to cg_get_attribute and convert to uint64_t
A common pattern in the codebase is reading a cgroup memory value
and converting it to a uint64_t. Let's make it a helper and refactor a
few places to use it so it's more concise.
Dan Streetman [Mon, 23 Mar 2020 21:34:17 +0000 (17:34 -0400)]
network: attach sd-event in link_load() when creating link dhcp_client or ipv4ll
Commit 08c588d18b1e337f856e6541f3f711be48718279 moved attachment of the
sd-event into creation of the dhcp_client or ipv4ll in dhcp4_configure()
or ipv4ll_configure(), but these can also be created in link_load(),
so that creation needs to also perform sd-event attachment.
Without this, dhcp_client or ipv4ll created in link_load() will not have
an ->event and will fail assertion, causing networkd to fail, e.g.:
Assertion 'client->event' failed at src/libsystemd-network/sd-dhcp-client.c:1283, function client_start_delayed(). Ignoring.
ens2: Could not acquire DHCPv4 lease: Invalid argument
ens2: Failed
Arusekk [Sun, 22 Mar 2020 09:41:18 +0000 (10:41 +0100)]
Fall back to kexec when no kexec binary exists
This may be not a popular setup, but in case kexec resides somewhere different
than the default location for KEXEC (`/usr/sbin/kexec`), don't just reboot and
try doing `reboot(RB_KEXEC)` instead, just like what `kexec -e` normally does.
Mar 21 23:26:32 Zeus systemd-networkd[671]: veth99: Link DOWN
Mar 21 23:26:32 Zeus systemd-networkd[671]: veth99: Lost carrier
Mar 22 10:25:35 Zeus systemd-networkd[671]: veth99: Link UP
Mar 22 10:25:35 Zeus systemd-networkd[671]: veth99: Gained carrier
Mar 22 10:25:36 Zeus systemd-networkd[671]: veth99: Gained IPv6LL
Mar 22 10:29:02 Zeus systemd-networkd[671]: veth99: Link DOWN
Mar 22 10:29:02 Zeus systemd-networkd[671]: veth99: Lost carrier
Mar 22 10:29:13 Zeus systemd-networkd[671]: veth99: Link UP
Mar 22 10:29:13 Zeus systemd-networkd[671]: veth99: Gained carrier
Mar 22 10:29:14 Zeus systemd-networkd[671]: veth99: Gained IPv6LL
sd-bus: support SD_BUS_DEFAULT* and don't crash in functions where bus is optional
In those functions where bus defaults to the m->bus, we should also
resolve the magic parameters. And if neither called with bus=NULL
and an unattached message, return properly instead of crashing in assert
later.
sd-bus: make sure SD_BUS_DEFAULT* works everywhere
I'm not sure why those functions were not touched in 45b1f410ba699990e4033805fe64cc369b835f10. Anyway, it seems easier
to just support the magic parameters everywhere.
Arnaud Ferraris [Sat, 14 Mar 2020 01:08:50 +0000 (02:08 +0100)]
login: allow non-console sessions to change vt
When starting a wayland session through a systemd service for a non-root
user, the compositor (based on wlroots) is denied the authorization to
change vt.
Once the user logs in, either through a local console or via ssh, the
compositor can work properly.
This is related to the login polkit policy:
- `allow_inactive` has value `auth_admin_keep`, denying any non-root
user session the authorization to change vt
- `allow_active` has value `yes`, which explains why the vt change
becomes possible once the user logs in through another channel
By changing the `allow_inactive` value to `yes`, any user session setup
in a service file can switch vt, allowing wayland sessions for non-root
users.
David Wood [Fri, 28 Feb 2020 18:28:49 +0000 (18:28 +0000)]
network: Fix split in `SendOption=` on client and server
When specifying `DHCPv4.SendOption=`, it is used by systemd-networkd to
set the value of that option within the DHCP request that is sent out.
This differs to setting `DHCPServer.SendOption=`, which will place all
the options together as suboptions into the vendor-specific information
(code 43) option.
This commit adds two new config options, `DHCPv4.SendVendorOption=` and
`DHCPServer.SendVendorOption=`. These both have the behaviour of the old
`DHCPServer.SendOption=` flag, and set the value of the suboption in the
vendor-specific information option.
The behaviour of `DHCPServer.SendOption=` is then changed to reflect
that of `DHCPv4.SendOption=`. It will set the value of the corresponding
option in the DHCP request.
selinux: delay mac_selinux_enforcing call after SELinux was determined to be enabled
Calling `mac_selinux_enforcing()`, which calls `security_getenforce()`, on a SELinux disabled system causes the following error message to be printed:
Failed to get SELinux enforced status: No such file or directory
Fixes: 257188f80ce1a083e3a88b679b898a73fecab53b ("selinux: cache enforced status and treat retrieve failure as enforced mode")
Supersedes: #15145
With the default config repeated press of fn+VolumeUp or fn+VolumeDown lead to
repeated presses even after release, which leads to volume going to 100% or 0%.
core: consistently emit oom warning when parsing args
In practice we are very unlikely to fail at this point, but for
consistency, we should always warn when allocation fails, and
we have free_and_strdup_warn() for this.