polkit: when authorizing via PK let's re-resolve callback/userdata instead of caching it
Previously, when doing an async PK query we'd store the original
callback/userdata pair and call it again after the PK request is
complete. This is problematic, since PK queries might be slow and in the
meantime the userdata might be released and re-acquired. Let's avoid
this by always traversing through the message handlers so that we always
re-resolve the callback and userdata pair and thus can be sure it's
up-to-date and properly valid.
sd-bus: introduce API for re-enqueuing incoming messages
When authorizing via PolicyKit we want to process incoming method calls
twice: once to process and figure out that we need PK authentication,
and a second time after we acquired PK authentication to actually execute
the operation. With this new call sd_bus_enqueue_for_read() we have a
way to put an incoming message back into the read queue for this
purpose.
This might have other uses too, for example debugging.
polkit: on async pk requests, re-validate action/details
When we do an async pk request, let's store which action/details we used
for the original request, and when we are called for the second time,
let's compare. If the action/details changed, let's not allow the access
to go through.
core: set fs.file-max sysctl to LONG_MAX rather than ULONG_MAX
Since kernel 5.2 the kernel thankfully returns proper errors when we
write a value out of range to the sysctl. Which however breaks writing
ULONG_MAX to request the maximum value. Hence let's write the new
maximum value instead, LONG_MAX.
Tom Yan [Wed, 9 Jan 2019 15:35:24 +0000 (23:35 +0800)]
mount/generators: do not make unit wanted by its device unit
As device units will be reloaded by systemd whenever the corresponding device generates a "changed" event, if the mount unit / cryptsetup service is wanted by its device unit, the former can be restarted by systemd unexpectedly after the user stopped them explicitly. It is not sensible at all and can be considered dangerous. Neither is the behaviour conventional (as `auto` in fstab should only affect behaviour on boot and `mount -a`) or ever documented at all (not even in systemd, see systemd.mount(5) and crypttab(5)).
Franck Bui [Wed, 13 Feb 2019 17:45:36 +0000 (18:45 +0100)]
namespace: make MountFlags=shared work again
Since commit 0722b359342d2a9f9e0d453875624387a0ba1be2, the root mountpoint is
unconditionnally turned to slave which breaks units that are using explicitly
MountFlags=shared (and no other options that would implicitly require a slave
root mountpoint).
This only affects systemd-resolved. bus_open_system_watch_bind_with_description()
is also used in timesyncd, but it has no methods, only read-only properties, and
in networkd, but it annotates all methods with SD_BUS_VTABLE_UNPRIVILEGED and does
polkit checks.
Michael Biebl [Wed, 17 Jul 2019 23:24:00 +0000 (01:24 +0200)]
meson: make nologin path build time configurable
Some distros install nologin as /usr/sbin/nologin, others as
/sbin/nologin.
Since we can't really on merged-usr everywhere (where the path wouldn't
matter), make the path build time configurable via -Dnologin-path=.
Susant Sahani [Thu, 9 May 2019 02:05:35 +0000 (07:35 +0530)]
networkd: fix link_up() (#12505)
Fillup IFLA_INET6_ADDR_GEN_MODE while we do link_up.
Fixes the following error:
```
dummy-test: Could not bring up interface: Invalid argument
```
After reading the kernel code when we do a link up
```
net/core/rtnetlink.c
IFLA_AF_SPEC
af_ops->set_link_af(dev, af);
inet6_set_link_af
if (tb[IFLA_INET6_ADDR_GEN_MODE])
Here it looks for IFLA_INET6_ADDR_GEN_MODE
```
Since link up we didn't filling up that it's failing.
Michael Biebl [Thu, 27 Jun 2019 13:02:40 +0000 (15:02 +0200)]
ask-password: prevent buffer overflow when reading from keyring
When we read from keyring, a temporary buffer is allocated in order to
determine the size needed for the entire data. However, when zeroing that area,
we use the data size returned by the read instead of the lesser size allocate
for the buffer.
That will cause memory corruption that causes systemd-cryptsetup to crash
either when a single large password is used or when multiple passwords have
already been pushed to the keyring.
Michael Biebl [Tue, 14 May 2019 11:12:35 +0000 (13:12 +0200)]
random-util: eat up bad RDRAND values seen on AMD CPUs
An ugly, ugly work-around for #11810. And no, we shouldn't have to do
this. This is something for AMD, the firmware or the kernel to
fix/work-around, not us. But nonetheless, this should do it for now.
Commit d85515edcf9700dc068201ab9f7103f04f3b25b2 changed logic how reboot is
executed. That commit changed behavior to use emergency action reboot code path
to perform the reboot.
This inadvertently broke rebooting with argument:
$ systemctl reboot custom-reason
Restore original behavior so that if reboot service unit similar to
systemd-reboot.service is executed it is possible to override reboot reason
with "systemctl reboot ARG".
When "systemctl reboot ARG" is executed ARG is placed in file
/run/systemd/reboot-param and reboot is issued using logind's Reboot
dbus-service.
If RebootArgument is specified in systemd-reboot.service it takes precedence
over what systemctl sets.
Yu Watanabe [Fri, 22 Feb 2019 04:27:44 +0000 (13:27 +0900)]
network: remove routing policy rule from foreign rule database when it is removed
Previously, When the first link configures rules, it removes all saved
rules, which were configured by networkd previously, in the foreign rule
database, but the rules themselves are still in the database.
Thus, when the second or later link configures rules, it errnously
treats the rules already exist.
This is the root of issue #11280.
This removes rules from the foreign database when they are removed.
Previously, RUN and SECLABEL keys are stored in udev_list with its unique
flag is false. If the flag is false, then udev_list is just a linked
list and new entries are always added in the last.
So, we should use OrderedHashmap instead of Hashmap.
remove "." path components from required mount paths
unit_require_mounts_for may be passed path arguments that contain "."
components like for user's home directories where "." is sometimes used
to specify some form of anchor point.
This change stops considering such path as an error and removes the "."
components instead.
cgtop: Fix processing of controllers other than CPU
After debugging the issue with gdb, I found that the following change
94ddb08 "cgtop: Still try to get CPU statistics if controller-free"
has introduced a bug, which prevents process(..) method processing
memory and io controllers when cpu_accounting_is_cheap() is true.
The obvious fix is to move this branch to be the last one, keeping
the intended behavior of the above change, without having a negative
effect on the other controllers.
Fixes #11773 [systemd-cgtop no longer shows memory (and io) usage]
Martin Pitt [Wed, 27 Feb 2019 22:15:31 +0000 (23:15 +0100)]
networkd-test: ignore failures of test_route_only_dns* in containers
This test exposes a race condition when running in LXC, see issue #11848
for details. Until that is understood and fixed, skip the test as it's
not a recent regression.
Martin Pitt [Thu, 21 Feb 2019 11:34:23 +0000 (12:34 +0100)]
networkd-test: fix test_dropin()
- This test needs resolved, so make sure it is started. In some Debian
environments it is not.
- It was an unnecessary, and now failing assumption that name servers
get atomically written to the resolved's resolv.conf. Wait until both
expected name servers are in the file.
Martin Pitt [Thu, 21 Feb 2019 11:26:44 +0000 (12:26 +0100)]
networkd-test: use a complete domain name in test_route_only_dns()
Since version 241 (commit ea4678?), querying MX type records for
single-label domains does not actually forward the query to the DNS
server any more. Use "example.com" instead, which is the recommended
test domain anyway.
Martin Pitt [Thu, 21 Feb 2019 11:24:16 +0000 (12:24 +0100)]
networkd-test: disable DNSSEC in domain-restricted DNS test
dnsmasq 2.80 changed behaviour when being queried by resolved with
enabled DNSSEC: It returns errors for SOA and DS queries which cause the
entire query to fail. As we don't configure DNSSEC in this test anyway,
just disable it so that we retain compatibility with old and new dnsmasq
versions.
Before a3ce813697bcc1c4644e097a2f1cd0459326d6ee, all properties in
src->properties and src->properties_db are mixed and copied to
dst->properties_db by device_copy_properties().
So, it is not necessary to store data from udev database file to
sd_device::properties_db before copying properties.
But now, properties are not mixed. So, the read data need to be
stored to also ::properties_db.
Martin Pitt [Sun, 17 Feb 2019 09:17:45 +0000 (10:17 +0100)]
sd-bus: enforce a size limit on D-Bus object paths
Replace stack with heap allocation. This avoids accessing/modifying
memory outside of the allocated stack region by sending specially
crafted D-Bus messages with very large object paths.
Vulnerability discovered by Chris Coulson <chris.coulson@canonical.com>,
patch provided by Riccardo Schirone <rschiron@redhat.com>.
CVE-2019-6454
Gbp-Pq: Name sd-bus-enforce-a-size-limit-on-D-Bus-object-paths.patch
Michael Biebl [Thu, 18 Mar 2021 19:59:14 +0000 (19:59 +0000)]
systemd (241-7~deb10u7) buster; urgency=medium
* core: make sure to restore the control command id, too.
Fixes a segfault in systemd that can be triggered when both
daemon-reload and a service restart happen concurrently. (Closes: #984495)
* seccomp: allow turning off of seccomp filtering via env var.
Since glibc 2.33 faccessat() is implemented via faccessat2(), which
is breaking running containers that use such a version of glibc under
systemd-nspawn in Buster.
Turning off seccomp filtering via the SYSTEMD_SECCOMP env var makes it
possible to run such new containers. (Closes: #984573)