shell-completion: udevadm: support --uuid option

Follow-up for 730b9c1e1469dfd6d6850e9ea06da3cac469eba0.

NEWS

@@ -5,9 +5,9 @@ CHANGES WITH 249 in spe:
         * When operating on disk images via the --image= switch of various
           tools (such as systemd-nspawn or systemd-dissect), and multiple
           suitable root or /usr/ partitions exist in the image, then a simple
-          strverscmp() inspired comparison is done on the GPT partition label,
-          and the newest partition picked. This permits a simple and generic
-          whole-file-system A/B update logic where new operating system
+          comparison inspired by strverscmp() is done on the GPT partition
+          label, and the newest partition is picked. This permits a simple and
+          generic whole-file-system A/B update logic where new operating system
           versions are dropped into partitions whose label is then updated with
           a matching version identifier.
 
@@ -34,9 +34,6 @@ CHANGES WITH 249 in spe:
           previously unprovisioned images (i.e. images with a mostly empty
           /etc/).
 
-        * The systemd-ask-password now also supports reading passwords from the
-          credentials subsystem, via the new --credential= switch.
-
         * Services gained a new ExitType= setting which can configure how to
           determine when a service exited: the default is "main" which defines
           the runtime by the service's main process lifetime (this matches the
@@ -47,30 +44,31 @@ CHANGES WITH 249 in spe:
         * The systemd-machine-id-setup tool now supports a --image= switch for
           provisioning a machine ID file into an OS disk image, similar to how
           --root= operates on an OS file tree. This matches the existing switch
-          of the same name to the systemd-tmpfiles, systemd-firstboot or
+          of the same name for systemd-tmpfiles, systemd-firstboot, and
           systemd-sysusers tools.
 
-        * Similar, systemd-repart gained support for a --image= switch, too. In
-          combination with the existing --size= makes the tool particularly
-          useful for easily growing disk images in a single command invocation,
-          following the declarative rules included in the image itself.
+        * Similarly, systemd-repart gained support for the --image= switch too.
+          In combination with the existing --size= option, this makes the tool
+          particularly useful for easily growing disk images in a single
+          invocation, following the declarative rules included in the image
+          itself.
 
         * systemd-repart's partition configuration files gained support for a
           new switch MakeDirectories= which may be used to create arbitrary
           directories inside file systems that are created, before registering
-          them in the partition table. This is useful in particular if root
-          partitions are created that way to create mount point directories for
-          other partitions included in the image. For example, a disk image
-          that is created to contain a /home/ and a /var/ partition in addition
-          to a root partition may MakeDirectories= to create /home/ and /var/
-          as empty directories in the root file system on creation so that the
-          resulting image can mounted immediately, even in read-only mode.
+          them in the partition table. This is useful in particular for root
+          partitions to create mount point directories for other partitions
+          included in the image. For example, a disk image that contains a
+          root, /home/, and /var/ partitions, may set MakeDirectories=yes to
+          create /home/ and /var/ as empty directories in the root file system
+          on its creation, so that the resulting image can mounted immediately,
+          even in read-only mode.
 
-        * systemd-repart's CopyBlocks= setting gained support for a special
-          value "auto". If used a suitable, matching partition on the booted OS
-          is found as source to copy blocks from. This is useful for
+        * systemd-repart's CopyBlocks= setting gained support for the special
+          value "auto". If used, a suitable matching partition on the booted OS
+          is found as source to copy blocks from. This is useful when
           implementing replicating installers, that are booted from one medium
-          and then stream their own root partition on the target medium.
+          and then stream their own root partition onto the target medium.
 
         * systemd-repart's partition configuration files gained support for a
           Flags= and a ReadOnly= setting, allowing control of the GPT partition
@@ -78,36 +76,54 @@ CHANGES WITH 249 in spe:
           created partitions as read-only from the start.
 
         * The /etc/os-release file has been extended with two new (optional)
-          variables IMAGE_VERSION= and IMAGE_ID=, for carrying identification
-          and versioning information for OS images that are updated
-          consistently, comprehensively and atomically as one image. The two
-          new specifiers %M, %A now resolve to these two fields in the various
-          configuration options that resolve specifiers.
+          variables IMAGE_VERSION= and IMAGE_ID=, carrying identity and version
+          information for OS images that are updated comprehensively and
+          atomically as one image. Two new specifiers %M, %A now resolve to
+          these two fields in the various configuration options that resolve
+          specifiers.
 
         * portablectl gained a new switch --extension= for enabling portable
           service images with extensions that follow the extension image
           concept introduced with v248.
 
         * systemd-coredump will now extract ELF build-id information from
-          processes dumping core and include it in the coredump
-          report. Moreover, if will look for ELF .note.package sections that
-          may carry distribution packaging meta-information about the crashing
-          process. This is useful to directly embed the RPM or Debian (or any
-          other) package name and version in the ELF files where they originate
-          from, making it easy to match up coredump reports with the software
-          versions they were found it. This is particular useful on
-          environments with ELF files form multiple vendors, different
-          distributions and versions, as it is common today, in particular in a
-          containerized and sand-boxed world. For further information, see:
+          processes dumping core and include it in the coredump report.
+          Moreover, it will look for ELF .note.package sections with
+          distribution packaging meta-information about the crashing process.
+          This is useful to directly embed the rpm or deb (or any other)
+          package name and version in ELF files, making it easy to match
+          coredump reports with the specific package for which the software was
+          compiled. This is particularly useful on environments with ELF files
+          from multiple vendors, different distributions and versions, as is
+          common today in our containerized and sand-boxed world. For further
+          information, see:
 
           https://systemd.io/COREDUMP_PACKAGE_METADATA
 
-        * A new udev hwdb has been added for Firewire audio devices
+        * A new udev hardware database has been added for FireWire devices
+          (IEEE 1394).
 
         * sd-bus' sd_bus_is_ready() and sd_bus_is_open() calls now accept a
           NULL bus object, for which they will return false. Or in other words,
           an unallocated bus connection is neither ready nor open.
 
+        * The sd-device API acquired a new API function
+          sd_device_get_usec_initialized() that returns the monotonic time when
+          the udev device first appeared in the database.
+
+        * sd-device gained a new APIs sd_device_trigger_with_uuid() and
+          sd_device_get_trigger_uuid(). The former is similar to
+          sd_device_trigger() but returns a randomly generated UUID that is
+          associated with the synthetic uevent generated by the call. This UUID
+          may be read from the sd_device object a monitor eventually receives,
+          via the sd_device_get_trigger_uuid(). This interface requires kernel
+          4.13 or above to work, and allows tracking a synthetic uevent through
+          the entire device management stack. The "udevadm trigger --settle"
+          logic has been updated to make use of this concept if available to
+          wait precisely for the uevents it generates. "udevadm trigger" also
+          gained a new parameter --uuid that prints the UUID for each generated
+          uevent.
+
         * The native Journal protocol has been documented. Clients may talk
           this as alternative to the classic BSD syslog protocol for locally
           delivering log records to the Journal. The protocol has been stable
@@ -142,81 +158,90 @@ CHANGES WITH 249 in spe:
           fine-tune whether to require an IPv4 or IPv6 address in order to
           consider an interface "online".
 
+        * networkctl will now show an over-all "online" state in the per-link
+          information.
+
+        * In .network files a new OutgoingInterface= setting has been added to
+          specify the output interface in bridge FDB setups.
+
+        * In ,network files the Multipath group ID may now be configured for
+          [NextHop] entries, via the new Group= setting.
+
+        * The DHCP server logic configured in .network files gained a new
+          setting RelayTarget= that turns the server into a DHCP server relay.
+          The RelayAgentCircuitId= and RelayAgentRemoteId= settings may be used
+          to further tweak the DHCP relay behaviour.
+
+        * The DHCP server logic also gained a new ServerAddress= setting in
+          .network files that explicitly specifies the server IP address to
+          use. If not specified, the address is determined automatically, as
+          before.
+
+        * The DHCP server logic in systemd-networkd gained support for static
+          DHCP leases, configurable via the [DHCPServerStaticLease]
+          section. This allows explicitly mapping specific MAC addresses to
+          fixed IP addresses and vice versa.
+
         * The RestrictAddressFamilies= setting in service files now supports a
           new special value "none". If specified sockets of all address
           families will be made unavailable to services configured that way.
 
-        * The DHCP server logic configured in .network files gained a new
-          setting RelayTarget= that turns the server into a DHCP server
-          relay. The RelayAgentCircuitId= and RelayAgentRemoteId= settings may
-          be used to further tweak the DHCP relay behaviour.
-
-        * The DHCP server logic also gained a new setting ServerAddress= in
-          .network files for explicitly specifying the server IP address to
-          use. If not used, the address is determined automatically, as before.
-
-        * The sd-device API acquired a new API function
-          sd_device_get_usec_initialized() that returns the monotonic timestamp
-          when a udev device appeared in the database first.
-
         * systemd-fstab-generator and systemd-repart have been updated to
           support booting from disks that carry only a /usr/ partition but no
-          root partition yet, but where systemd-repart can add them in on first
-          boot. This is useful for implementing systems that ship with a single
-          /usr/ file system and whose root file system shall be set up on a
-          LUKS encrypted volume whose key is generated locally (and possibly
-          enrolled in TPM), and which is set up and formatted during first
-          boot.
+          root partition yet, and where systemd-repart can add it in on the
+          first boot. This is useful for implementing systems that ship with a
+          single /usr/ file system, and whose root file system shall be set up
+          and formatted on a LUKS-encrypted volume whose key is generated
+          locally (and possibly enrolled in the TPM) during the first boot.
 
         * The [Address] section of .network files now accepts a new
           RouteMetric= setting that configures the routing metric to use for
-          the prefix route created as effect of the address
-          configuration. Similar, the [DHCPv6PrefixDelegation] and [IPv6Prefix]
-          sections gained matching settings for their prefix routes. (The
-          option of the same name of the [DHCPv6] section is moved to
-          [IPv6AcceptRA], since it conceptually belongs there; the old option
-          is still understood for compatibility.)
+          the prefix route created as effect of the address configuration.
+          Similarly, the [DHCPv6PrefixDelegation] and [IPv6Prefix] sections
+          gained matching settings for their prefix routes. (The option of the
+          same name in the [DHCPv6] section is moved to [IPv6AcceptRA], since
+          it conceptually belongs there; the old option is still understood for
+          compatibility.)
 
         * The DHCPv6 IAID and DUID are now explicitly configurable in .network
           files.
 
         * A new udev property ID_NET_DHCP_BROADCAST on network interface
-          devices is not honoured by systemd-networkd, that controls whether to
+          devices is now honoured by systemd-networkd, controlling whether to
           issue DHCP offers via broadcasting. This is used to ensure that s390
           layer 3 network interfaces work out-of-the-box with systemd-networkd.
 
         * nss-myhostname and systemd-resolved will now synthesize address
           records for a new special hostname "_outbound". The name will always
           resolve to the local IP addresses most likely used for outbound
-          connections towards the default route. On multi-homed hosts this is
+          connections towards the default routes. On multi-homed hosts this is
           useful to have a stable handle referring to "the" local IP address
           that matters most, to the point where this is defined.
 
         * The Discoverable Partition Specification has been updated with a new
-          GPT partition flag defined its partition types: a "growfs"
-          flag. Whenever partitions with this flag set are automatically
-          mounted (i.e. via systemd-gpt-auto-generator or the --image= switch
-          of systemd-nspawn or other tools; and as opposed to explicit mounting
+          GPT partition flag "growsfs" defined for its partition types.
+          Whenever partitions with this flag set are automatically mounted
+          (i.e. via systemd-gpt-auto-generator or the --image= switch of
+          systemd-nspawn or other tools; and as opposed to explicit mounting
           via /etc/fstab), the file system within the partition is
-          automatically grown to the full size of the partition should it be
-          smaller. If the file system size already matches the partition size
-          this flag has no effect. Previously, this functionality has been
-          available via the explicit x-systemd.growfs mount option, and this
-          new flag extends this to automatically discovered mounts. A new
-          GrowFileSystem= setting has been added to systemd-repart drop-in
-          files that allows configuring this partition flag. This new flag
-          defaults to on for partitions automatically created by
-          systemd-repart, except if they are marked read-only. See the
-          specification for further details:
+          automatically grown to the full size of the partition. If the file
+          system size already matches the partition size this flag has no
+          effect. Previously, this functionality has been available via the
+          explicit x-systemd.growfs mount option, and this new flag extends
+          this to automatically discovered mounts. A new GrowFileSystem=
+          setting has been added to systemd-repart drop-in files that allows
+          configuring this partition flag. This new flag defaults to on for
+          partitions automatically created by systemd-repart, except if they
+          are marked read-only. See the specification for further details:
 
           https://systemd.io/DISCOVERABLE_PARTITIONS
 
-        * .network files gained a new setting RoutToNTP= in the [DHCPv4]
+        * .network files gained a new setting RoutesToNTP= in the [DHCPv4]
           section. If enabled (which is the default), and an NTP server address
           is acquired through a DHCP lease on this interface an explicit route
           to this address is created on this interface to ensure that NTP
           traffic to the NTP server acquired on an interface is also routed
-          trough that interface. The pre-existing RoutesToDNS= setting that
+          through that interface. The pre-existing RoutesToDNS= setting that
           implements the same for DNS servers is now enabled by default.
 
         * A pair of service settings SocketBindAllow= + SocketBindDeny= have
@@ -224,27 +249,27 @@ CHANGES WITH 249 in spe:
           sockets created by the service may be bound to. This is implemented
           via BPF.
 
-        * A new ConditionFirmware= condition type has been added to unit
-          files. It may be used to detect certain firmware features. At the
-          moment it may check whether running on an UEFI system, a device.tree
-          system or if the system is compatible with some specified device-tree
-          feature.
+        * A new ConditionFirmware= condition type has been added to unit files.
+          It may be used to detect certain firmware features. At the moment it
+          may check whether running on an UEFI system, a device.tree system, or
+          if the system is compatible with some specified device-tree feature.
 
         * hostnamed gained a new Describe() D-Bus method that returns a JSON
           serialization of the host data it exposes. This is exposed via
-          "hostnamectl --json=" to acquire a host identity description in
-          JSON. It's our intention to add a similar concept to most services
-          and objects systemd manages, in order to simplify integration with
-          program code the consumes JSON.
+          "hostnamectl --json=" to acquire a host identity description in JSON.
+          It's our intention to add a similar features to most services and
+          objects systemd manages, in order to simplify integration with
+          program code that can consume JSON.
 
-        * Similar, networkd gained a Describe() method on its Manager and Link
-          bus objects. This is exposed via "networkctl --json=".
+        * Similarly, networkd gained a Describe() method on its Manager and
+          Link bus objects. This is exposed via "networkctl --json=".
 
-        * hostnamectl's various set-xyz verbs (e.g. "hostnamectl set-hostname")
-          have been renamed to just xyz (e.g. "hostnamectl hostname") and may
-          now be used to print the indicated data in terse form, instead of
-          only setting it. The old names continue to be supported for
-          compatibility.
+        * hostnamectl's various "get-xyz"/"set-xyz" verb pairs
+          (e.g. "hostnamectl get-hostname", "hostnamectl "set-hostname") have
+          been replaced by a single "xyz" verb (e.g. "hostnamectl hostname")
+          that is used both to get the value (when no argument is given), and
+          to set the value (when an argument is specified). The old names
+          continue to be supported for compatibility.
 
         * systemd-detect-virt and ConditionVirtualization= are now able to
           correctly identify Amazon EC2 environments.
@@ -257,8 +282,8 @@ CHANGES WITH 249 in spe:
 
         * bootctl gained support for a new --make-machine-id-directory= switch
           that allows precise control on whether to create the top-level
-          per-machine directory in the boot partition that typically contain
-          Type #1 boot loader entries.
+          per-machine directory in the boot partition that typically contains
+          Type 1 boot loader entries.
 
         * During build SBAT data to include in the systemd-boot EFI PE binaries
           may be specified now.
@@ -344,15 +369,6 @@ CHANGES WITH 249 in spe:
           journalctl, and permit exposing only the specified subset of the
           Journal records.
 
-        * networkctl will now show an over-all "online" state in the per-link
-          information.
-
-        * In .network files a new OutgoingInterface= setting has been added to
-          specify the output interface in bridge FDB setups.
-
-        * In ,network files the Multipath group ID may now be configured for
-          [NextHop] entries, via the new Group= setting.
-
         * The OnFailure= dependency between units is now augmented with a
           implicit reverse dependency OnFailureOf= (this new dependency cannot
           be configured directly it's only created as effect of an OnFailure=
@@ -366,6 +382,46 @@ CHANGES WITH 249 in spe:
           stop events between two units. It operates similar to the existing
           PropagatesReloadTo= + ReloadPropagatedFrom= dependencies.
 
+        * A new dependency type OnSuccess= has been added (plus the reverse
+          dependency OnSuccessOf=, which cannot be configured directly, but
+          exists only as effect of the reverse OnSuccess=). It is similar to
+          OnFailure=, but triggers in the opposite case: when a service exits
+          cleanly. This allows "chaining up" of services where one or more
+          services are started once another service has successfully completed.
+
+        * A new dependency type Upholds= has been added (plus the reverse
+          dependency UpheldBy=, which cannot be configured directly, but exists
+          only as effect of Upholds=). This dependency type is a stronger form
+          of Wants=: if a unit has an UpHolds= dependency on some other unit
+          and the former is active then the latter is started whenever it is
+          found inactive (and no job is queued for it). This is an alternative
+          to Restart= inside service units, but less configurable, and the
+          request to uphold a unit is not encoded in the unit itself but in
+          another unit that intends to uphold it.
+
+        * The systemd-ask-password tool now also supports reading passwords
+          from the credentials subsystem, via the new --credential= switch.
+
+        * The systemd-ask-password tool learnt a new switch --emoji= which may
+          be used to explicit control whether the lock and key emoji (🔐) is
+          shown in the password prompt on suitable TTYs.
+
+        * The --echo switch of systemd-ask-password now optionally takes a
+          parameter that controls character echo. It may either show asterisks
+          (default, as before), turn echo off entirely, or echo the typed
+          characters literally.
+
+        * New documentation has been added that describes the organization of
+          the systemd source code tree:
+
+          https://systemd.io/ARCHITECTURE
+
+        * …
+
+        Contributions from: …
+
+        — Berlin, 2021-06-XX
+
 CHANGES WITH 248:
 
         * A concept of system extension images is introduced. Such images may

docs/ARCHITECTURE.md

@@ -1,5 +1,5 @@
 ---
-title: systemd repository architecture
+title: systemd Repository Architecture
 category: Contributing
 layout: default
 ---
@@ -13,19 +13,19 @@ components of the systemd repository.
 
 Directories in `src/` provide the implementation of all daemons, libraries and
 command-line tools shipped by the project. There are many, and more are
-constantly added, so we will not enumerate them all here - the directory
+constantly added, so we will not enumerate them all here — the directory
 names are self-explanatory.
 
 ## Shared Code
 
 You might wonder what kind of common code belongs in `src/shared/` and what
 belongs in `src/basic/`. The split is like this: anything that is used to
-implement the public shared object we provide (sd-bus, sd-login, sd-id128,
-nss-systemd, nss-mymachines, nss-resolve, nss-myhostname, pam_systemd), must
-be located in `src/basic` (those objects are not allowed to link to
-libsystemd-shared.so). Conversely, anything which is shared between multiple
-components and does not need to be in `src/basic/`, should be in
-`src/shared/`.
+implement the public shared objects we provide (`sd-bus`, `sd-login`,
+`sd-id128`, `nss-systemd`, `nss-mymachines`, `nss-resolve`, `nss-myhostname`,
+`pam_systemd`), must be located in `src/basic` (those objects are not allowed
+to link to `libsystemd-shared.so`). Conversely, anything which is shared
+between multiple components and does not need to be in `src/basic/`, should be
+in `src/shared/`.
 
 To summarize:
 
@@ -63,10 +63,10 @@ file is compiled in a standalone binary that can be run to exercise the
 corresponding module. While most of the tests can be ran by any user, some
 require privileges, and will attempt to clearly log about what they need
 (mostly in the form of effective capabilities). These tests are self-contained,
-and generally safe to run on a host without side effects.
+and generally safe to run on the host without side effects.
 
-Ideally, every module in `src/basic/` and `src/shared/` should have a corresponding
-unit test under `src/test/`, which exercises every helper function.
+Ideally, every module in `src/basic/` and `src/shared/` should have a
+corresponding unit test under `src/test/`, exercising every helper function.
 
 # Integration Tests
 

hwdb.d/60-keyboard.hwdb

@@ -625,9 +625,9 @@ evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHewlett-Packard*:pnHPEliteBook*:*
 evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHP*:pnHPEliteBook*:*
 # HP ProBook 440 G2
 evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHewlett-Packard*:pnHP440G2:*
-# several HP ProBooks 4xx
-evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHewlett-Packard*:pnHP*ProBook4*:*
-evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHP*:pnHP*ProBook*4*:*
+# HP ProBook
+evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHewlett-Packard*:pnHP*ProBook*:*
+evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHP*:pnHP*ProBook*:*
 # HP ZBook
 evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHewlett-Packard*:pnHPZBook*:*
 evdev:atkbd:dmi:bvn*:bvr*:bd*:svnHP*:pnHPZBook*:*

shell-completion/bash/udevadm

@@ -51,7 +51,7 @@ _udevadm() {
         [INFO_STANDALONE]='-r --root -a --attribute-walk -x --export -e --export-db -c --cleanup-db
                            -w --wait-for-initialization'
         [INFO_ARG]='-q --query -p --path -n --name -P --export-prefix -d --device-id-of-file'
-        [TRIGGER_STANDALONE]='-v --verbose -n --dry-run -q --quiet -w --settle --wait-daemon'
+        [TRIGGER_STANDALONE]='-v --verbose -n --dry-run -q --quiet -w --settle --wait-daemon --uuid'
         [TRIGGER_ARG]='-t --type -c --action -s --subsystem-match -S --subsystem-nomatch
                        -a --attr-match -A --attr-nomatch -p --property-match
                        -g --tag-match -y --sysname-match --name-match -b --parent-match'

shell-completion/zsh/_udevadm

@@ -31,7 +31,8 @@ _udevadm_trigger(){
         '--property-match=[Trigger events for devices with a matching property value.]' \
         '--tag-match=property[Trigger events for devices with a matching tag.]' \
         '--sysname-match=[Trigger events for devices with a matching sys device name.]' \
-        '--parent-match=[Trigger events for all children of a given device.]'
+        '--parent-match=[Trigger events for all children of a given device.]' \
+        '--uuid[Print synthetic uevent UUID.]'
 }
 
 (( $+functions[_udevadm_settle] )) ||

src/basic/btrfs-util.c

@@ -19,7 +19,6 @@
 #include "btrfs-util.h"
 #include "chattr-util.h"
 #include "copy.h"
-#include "device-nodes.h"
 #include "fd-util.h"
 #include "fileio.h"
 #include "fs-util.h"

src/basic/device-nodes.h

@@ -1,16 +0,0 @@
-/* SPDX-License-Identifier: LGPL-2.1-or-later */
-#pragma once
-
-#include <stddef.h>
-#include <sys/types.h>
-
-#include "macro.h"
-#include "stdio-util.h"
-
-int encode_devnode_name(const char *str, char *str_enc, size_t len);
-int allow_listed_char_for_devnode(char c, const char *additional);
-
-#define DEV_NUM_PATH_MAX                                                \
-        (STRLEN("/dev/block/") + DECIMAL_STR_MAX(dev_t) + 1 + DECIMAL_STR_MAX(dev_t))
-#define xsprintf_dev_num_path(buf, type, devno)                         \
-        xsprintf(buf, "/dev/%s/%u:%u", type, major(devno), minor(devno))

src/basic/hexdecoct.c

@@ -526,6 +526,16 @@ char base64char(int x) {
         return table[x & 63];
 }
 
+/* This is almost base64char(), but not entirely, as it uses the "url and filename safe" alphabet,
+ * since we don't want "/" appear in interface names (since interfaces appear in sysfs as filenames).
+ * See section #5 of RFC 4648. */
+char urlsafe_base64char(int x) {
+        static const char table[64] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+                                      "abcdefghijklmnopqrstuvwxyz"
+                                      "0123456789-_";
+        return table[x & 63];
+}
+
 int unbase64char(char c) {
         unsigned offset;
 

src/basic/hexdecoct.h

@@ -27,6 +27,7 @@ char base32hexchar(int x) _const_;
 int unbase32hexchar(char c) _const_;
 
 char base64char(int x) _const_;
+char urlsafe_base64char(int x) _const_;
 int unbase64char(char c) _const_;
 
 char *base32hexmem(const void *p, size_t l, bool padding);

src/basic/meson.build

@@ -38,8 +38,6 @@ basic_sources = files('''
         creds-util.c
         creds-util.h
         def.h
-        device-nodes.c
-        device-nodes.h
         dirent-util.c
         dirent-util.h
         dlfcn-util.c

src/core/unit.c

@@ -1475,7 +1475,7 @@ static int unit_add_mount_dependencies(Unit *u) {
                         r = unit_name_from_path(prefix, ".mount", &p);
                         if (IN_SET(r, -EINVAL, -ENAMETOOLONG))
                                 continue; /* If the path cannot be converted to a mount unit name, then it's
-                                           * not managable as a unit by systemd, and hence we don't need a
+                                           * not manageable as a unit by systemd, and hence we don't need a
                                            * dependency on it. Let's thus silently ignore the issue. */
                         if (r < 0)
                                 return r;

src/nspawn/nspawn-network.c

@@ -11,6 +11,7 @@
 
 #include "alloc-util.h"
 #include "ether-addr-util.h"
+#include "hexdecoct.h"
 #include "lockfile-util.h"
 #include "missing_network.h"
 #include "netif-naming-scheme.h"
@@ -200,16 +201,6 @@ static int add_veth(
         return 0;
 }
 
-/* This is almost base64char(), but not entirely, as it uses the "url and filename safe" alphabet, since we
- * don't want "/" appear in interface names (since interfaces appear in sysfs as filenames). See section #5
- * of RFC 4648. */
-static char urlsafe_base64char(int x) {
-        static const char table[64] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
-                                      "abcdefghijklmnopqrstuvwxyz"
-                                      "0123456789-_";
-        return table[x & 63];
-}
-
 static int shorten_ifname(char *ifname) {
         char new_ifname[IFNAMSIZ];
 

src/shared/device-nodes.h

@@ -0,0 +1,7 @@
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+#pragma once
+
+#include <stddef.h>
+
+int encode_devnode_name(const char *str, char *str_enc, size_t len);
+int allow_listed_char_for_devnode(char c, const char *additional);

src/shared/meson.build

@@ -80,6 +80,8 @@ shared_sources = files('''
         daemon-util.h
         dev-setup.c
         dev-setup.h
+        device-nodes.c
+        device-nodes.h
         devnode-acl.h
         discover-image.c
         discover-image.h

src/test/test-device-nodes.c

@@ -1,11 +1,11 @@
 /* SPDX-License-Identifier: LGPL-2.1-or-later */
 
+#include <stdio.h>
 #include <sys/types.h>
 
 #include "alloc-util.h"
 #include "device-nodes.h"
 #include "string-util.h"
-#include "util.h"
 
 /* helpers for test_encode_devnode_name */
 static char *do_encode_string(const char *in) {

src/test/test-util.c

@@ -316,7 +316,7 @@ static void test_raw_clone(void) {
 
         errno = 0;
         assert_se(raw_clone(CLONE_FS|CLONE_NEWNS) == -1);
-        assert_se(errno == EINVAL);
+        assert_se(errno == EINVAL || ERRNO_IS_PRIVILEGE(errno)); /* Certain container environments prohibit namespaces to us, don't fail in that case */
 }
 
 static void test_physical_memory(void) {

src/udev/meson.build

@@ -200,6 +200,12 @@ tests += [
          [threads,
           libacl]],
 
+        [['src/udev/test-udev-node.c'],
+         [libudevd_core,
+          libshared],
+         [threads,
+          libacl]],
+
         [['src/udev/test-udev-builtin.c'],
          [libudevd_core,
           libshared],

src/udev/test-udev-node.c

@@ -0,0 +1,44 @@
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#include "tests.h"
+#include "udev-node.h"
+
+static void test_udev_node_escape_path_one(const char *path, const char *expected) {
+        char buf[NAME_MAX+1];
+        size_t r;
+
+        r = udev_node_escape_path(path, buf, sizeof buf);
+        log_debug("udev_node_escape_path(%s) -> %s (expected: %s)", path, buf, expected);
+        assert_se(r == strlen(expected));
+        assert_se(streq(buf, expected));
+}
+
+static void test_udev_node_escape_path(void) {
+        char a[NAME_MAX+1], b[NAME_MAX+1];
+
+        test_udev_node_escape_path_one("/disk/by-id/nvme-eui.1922908022470001001b448b44ccb9d6", "\\x2fdisk\\x2fby-id\\x2fnvme-eui.1922908022470001001b448b44ccb9d6");
+        test_udev_node_escape_path_one("/disk/by-id/nvme-eui.1922908022470001001b448b44ccb9d6-part1", "\\x2fdisk\\x2fby-id\\x2fnvme-eui.1922908022470001001b448b44ccb9d6-part1");
+        test_udev_node_escape_path_one("/disk/by-id/nvme-eui.1922908022470001001b448b44ccb9d6-part2", "\\x2fdisk\\x2fby-id\\x2fnvme-eui.1922908022470001001b448b44ccb9d6-part2");
+        test_udev_node_escape_path_one("/disk/by-id/nvme-WDC_PC_SN720_SDAQNTW-512G-1001_192290802247", "\\x2fdisk\\x2fby-id\\x2fnvme-WDC_PC_SN720_SDAQNTW-512G-1001_192290802247");
+        test_udev_node_escape_path_one("/disk/by-id/nvme-WDC_PC_SN720_SDAQNTW-512G-1001_192290802247-part1", "\\x2fdisk\\x2fby-id\\x2fnvme-WDC_PC_SN720_SDAQNTW-512G-1001_192290802247-part1");
+        test_udev_node_escape_path_one("/disk/by-id/nvme-WDC_PC_SN720_SDAQNTW-512G-1001_192290802247-part2", "\\x2fdisk\\x2fby-id\\x2fnvme-WDC_PC_SN720_SDAQNTW-512G-1001_192290802247-part2");
+        test_udev_node_escape_path_one("/disk/by-id/usb-Generic-_SD_MMC_20120501030900000-0:0", "\\x2fdisk\\x2fby-id\\x2fusb-Generic-_SD_MMC_20120501030900000-0:0");
+
+        memset(a, 'a', sizeof(a) - 1);
+        memcpy(a, "/disk/by-id/", strlen("/disk/by-id/"));
+        char_array_0(a);
+
+        memset(b, 'a', sizeof(b) - 1);
+        memcpy(b, "\\x2fdisk\\x2fby-id\\x2f", strlen("\\x2fdisk\\x2fby-id\\x2f"));
+        strcpy(b + sizeof(b) - 12, "N3YhcCqFeID");
+
+        test_udev_node_escape_path_one(a, b);
+}
+
+int main(int argc, char *argv[]) {
+        test_setup_logging(LOG_INFO);
+
+        test_udev_node_escape_path();
+
+        return 0;
+}

src/udev/udev-node.c

@@ -7,14 +7,16 @@
 #include <sys/stat.h>
 #include <unistd.h>
 
+#include "sd-id128.h"
+
 #include "alloc-util.h"
-#include "device-nodes.h"
 #include "device-private.h"
 #include "device-util.h"
 #include "dirent-util.h"
 #include "fd-util.h"
 #include "format-util.h"
 #include "fs-util.h"
+#include "hexdecoct.h"
 #include "mkdir.h"
 #include "path-util.h"
 #include "selinux-util.h"
@@ -26,7 +28,36 @@
 #include "udev-node.h"
 #include "user-util.h"
 
+#define CREATE_LINK_MAX_RETRIES 128
 #define LINK_UPDATE_MAX_RETRIES 128
+#define TOUCH_FILE_MAX_RETRIES  128
+#define UDEV_NODE_HASH_KEY SD_ID128_MAKE(b9,6a,f1,ce,40,31,44,1a,9e,19,ec,8b,ae,f3,e3,2f)
+
+static int create_symlink(const char *target, const char *slink) {
+        int r;
+
+        assert(target);
+        assert(slink);
+
+        for (unsigned i = 0; i < CREATE_LINK_MAX_RETRIES; i++) {
+                r = mkdir_parents_label(slink, 0755);
+                if (r == -ENOENT)
+                        continue;
+                if (r < 0)
+                        return r;
+
+                mac_selinux_create_file_prepare(slink, S_IFLNK);
+                if (symlink(target, slink) < 0)
+                        r = -errno;
+                else
+                        r = 0;
+                mac_selinux_create_file_clear();
+                if (r != -ENOENT)
+                        return r;
+        }
+
+        return r;
+}
 
 static int node_symlink(sd_device *dev, const char *node, const char *slink) {
         _cleanup_free_ char *slink_dirname = NULL, *target = NULL;
@@ -38,87 +69,80 @@ static int node_symlink(sd_device *dev, const char *node, const char *slink) {
         assert(node);
         assert(slink);
 
-        slink_dirname = dirname_malloc(slink);
-        if (!slink_dirname)
-                return log_oom();
+        r = path_extract_directory(slink, &slink_dirname);
+        if (r < 0)
+                return log_device_debug_errno(dev, r, "Failed to get parent directory of '%s': %m", slink);
 
         /* use relative link */
         r = path_make_relative(slink_dirname, node, &target);
         if (r < 0)
-                return log_device_error_errno(dev, r, "Failed to get relative path from '%s' to '%s': %m", slink, node);
+                return log_device_debug_errno(dev, r, "Failed to get relative path from '%s' to '%s': %m", slink, node);
 
-        /* preserve link with correct target, do not replace node of other device */
-        if (lstat(slink, &stats) == 0) {
-                if (S_ISBLK(stats.st_mode) || S_ISCHR(stats.st_mode))
-                        return log_device_error_errno(dev, SYNTHETIC_ERRNO(EOPNOTSUPP),
-                                                      "Conflicting device node '%s' found, link to '%s' will not be created.", slink, node);
-                else if (S_ISLNK(stats.st_mode)) {
-                        _cleanup_free_ char *buf = NULL;
+        if (lstat(slink, &stats) >= 0) {
+                _cleanup_free_ char *buf = NULL;
+
+                if (!S_ISLNK(stats.st_mode))
+                        return log_device_debug_errno(dev, SYNTHETIC_ERRNO(EEXIST),
+                                                      "Conflicting inode '%s' found, link to '%s' will not be created.", slink, node);
+
+                if (readlink_malloc(slink, &buf) >= 0 &&
+                    path_equal(target, buf)) {
+                        /* preserve link with correct target, do not replace node of other device */
+                        log_device_debug(dev, "Preserve already existing symlink '%s' to '%s'", slink, target);
+
+                        (void) label_fix(slink, LABEL_IGNORE_ENOENT);
+                        (void) utimensat(AT_FDCWD, slink, NULL, AT_SYMLINK_NOFOLLOW);
 
-                        if (readlink_malloc(slink, &buf) >= 0 &&
-                            streq(target, buf)) {
-                                log_device_debug(dev, "Preserve already existing symlink '%s' to '%s'", slink, target);
-                                (void) label_fix(slink, LABEL_IGNORE_ENOENT);
-                                (void) utimensat(AT_FDCWD, slink, NULL, AT_SYMLINK_NOFOLLOW);
-                                return 0;
-                        }
-                }
-        } else {
-                log_device_debug(dev, "Creating symlink '%s' to '%s'", slink, target);
-                do {
-                        r = mkdir_parents_label(slink, 0755);
-                        if (!IN_SET(r, 0, -ENOENT))
-                                break;
-                        mac_selinux_create_file_prepare(slink, S_IFLNK);
-                        if (symlink(target, slink) < 0)
-                                r = -errno;
-                        mac_selinux_create_file_clear();
-                } while (r == -ENOENT);
-                if (r == 0)
                         return 0;
-                if (r < 0)
-                        log_device_debug_errno(dev, r, "Failed to create symlink '%s' to '%s', trying to replace '%s': %m", slink, target, slink);
-        }
+                }
+        } else if (errno == ENOENT) {
+                log_device_debug(dev, "Creating symlink '%s' to '%s'", slink, target);
+
+                r = create_symlink(target, slink);
+                if (r >= 0)
+                        return 0;
+
+                log_device_debug_errno(dev, r, "Failed to create symlink '%s' to '%s', trying to replace '%s': %m", slink, target, slink);
+        } else
+                return log_device_debug_errno(dev, errno, "Failed to lstat() '%s': %m", slink);
 
         log_device_debug(dev, "Atomically replace '%s'", slink);
+
         r = device_get_device_id(dev, &id);
         if (r < 0)
-                return log_device_error_errno(dev, r, "Failed to get device id: %m");
+                return log_device_debug_errno(dev, r, "Failed to get device id: %m");
         slink_tmp = strjoina(slink, ".tmp-", id);
+
         (void) unlink(slink_tmp);
-        do {
-                r = mkdir_parents_label(slink_tmp, 0755);
-                if (!IN_SET(r, 0, -ENOENT))
-                        break;
-                mac_selinux_create_file_prepare(slink_tmp, S_IFLNK);
-                if (symlink(target, slink_tmp) < 0)
-                        r = -errno;
-                mac_selinux_create_file_clear();
-        } while (r == -ENOENT);
+
+        r = create_symlink(target, slink_tmp);
         if (r < 0)
-                return log_device_error_errno(dev, r, "Failed to create symlink '%s' to '%s': %m", slink_tmp, target);
+                return log_device_debug_errno(dev, r, "Failed to create symlink '%s' to '%s': %m", slink_tmp, target);
 
         if (rename(slink_tmp, slink) < 0) {
-                r = log_device_error_errno(dev, errno, "Failed to rename '%s' to '%s': %m", slink_tmp, slink);
+                r = log_device_debug_errno(dev, errno, "Failed to rename '%s' to '%s': %m", slink_tmp, slink);
                 (void) unlink(slink_tmp);
-        } else
-                /* Tell caller that we replaced already existing symlink. */
-                r = 1;
+                return r;
+        }
 
-        return r;
+        /* Tell caller that we replaced already existing symlink. */
+        return 1;
 }
 
-/* find device node of device with highest priority */
 static int link_find_prioritized(sd_device *dev, bool add, const char *stackdir, char **ret) {
         _cleanup_closedir_ DIR *dir = NULL;
         _cleanup_free_ char *target = NULL;
         struct dirent *dent;
         int r, priority = 0;
+        const char *id;
 
-        assert(!add || dev);
+        assert(dev);
         assert(stackdir);
         assert(ret);
 
+        /* Find device node of device with highest priority. This returns 1 if a device found, 0 if no
+         * device found, or a negative errno. */
+
         if (add) {
                 const char *devnode;
 
@@ -137,17 +161,21 @@ static int link_find_prioritized(sd_device *dev, bool add, const char *stackdir,
 
         dir = opendir(stackdir);
         if (!dir) {
-                if (target) {
+                if (errno == ENOENT) {
                         *ret = TAKE_PTR(target);
-                        return 0;
+                        return !!*ret;
                 }
 
                 return -errno;
         }
 
+        r = device_get_device_id(dev, &id);
+        if (r < 0)
+                return r;
+
         FOREACH_DIRENT_ALL(dent, dir, break) {
                 _cleanup_(sd_device_unrefp) sd_device *dev_db = NULL;
-                const char *devnode, *id;
+                const char *devnode;
                 int db_prio = 0;
 
                 if (dent->d_name[0] == '\0')
@@ -157,9 +185,6 @@ static int link_find_prioritized(sd_device *dev, bool add, const char *stackdir,
 
                 log_device_debug(dev, "Found '%s' claiming '%s'", dent->d_name, stackdir);
 
-                if (device_get_device_id(dev, &id) < 0)
-                        continue;
-
                 /* did we find ourself? */
                 if (streq(dent->d_name, id))
                         continue;
@@ -184,59 +209,73 @@ static int link_find_prioritized(sd_device *dev, bool add, const char *stackdir,
                 priority = db_prio;
         }
 
-        if (!target)
-                return -ENOENT;
-
         *ret = TAKE_PTR(target);
-        return 0;
+        return !!*ret;
 }
 
-static size_t escape_path(const char *src, char *dest, size_t size) {
+size_t udev_node_escape_path(const char *src, char *dest, size_t size) {
         size_t i, j;
+        uint64_t h;
 
         assert(src);
         assert(dest);
 
         for (i = 0, j = 0; src[i] != '\0'; i++) {
                 if (src[i] == '/') {
-                        if (j+4 >= size) {
-                                j = 0;
-                                break;
-                        }
+                        if (j+4 >= size)
+                                goto toolong;
                         memcpy(&dest[j], "\\x2f", 4);
                         j += 4;
                 } else if (src[i] == '\\') {
-                        if (j+4 >= size) {
-                                j = 0;
-                                break;
-                        }
+                        if (j+4 >= size)
+                                goto toolong;
                         memcpy(&dest[j], "\\x5c", 4);
                         j += 4;
                 } else {
-                        if (j+1 >= size) {
-                                j = 0;
-                                break;
-                        }
+                        if (j+1 >= size)
+                                goto toolong;
                         dest[j] = src[i];
                         j++;
                 }
         }
         dest[j] = '\0';
         return j;
+
+toolong:
+        /* If the input path is too long to encode as a filename, then let's suffix with a string
+         * generated from the hash of the path. */
+
+        h = siphash24_string(src, UDEV_NODE_HASH_KEY.bytes);
+
+        assert(size >= 12);
+
+        for (unsigned k = 0; k <= 10; k++)
+                dest[size - k - 2] = urlsafe_base64char((h >> (k * 6)) & 63);
+
+        dest[size - 1] = '\0';
+        return size - 1;
 }
 
 /* manage "stack of names" with possibly specified device priorities */
-static int link_update(sd_device *dev, const char *slink, bool add) {
-        _cleanup_free_ char *filename = NULL, *dirname = NULL;
+static int link_update(sd_device *dev, const char *slink_in, bool add) {
+        _cleanup_free_ char *slink = NULL, *filename = NULL, *dirname = NULL;
         const char *slink_name, *id;
-        char name_enc[PATH_MAX];
+        char name_enc[NAME_MAX+1];
         int i, r, retries;
 
         assert(dev);
-        assert(slink);
+        assert(slink_in);
+
+        slink = strdup(slink_in);
+        if (!slink)
+                return log_oom_debug();
+
+        path_simplify(slink);
 
         slink_name = path_startswith(slink, "/dev");
-        if (!slink_name)
+        if (!slink_name ||
+            empty_or_root(slink_name) ||
+            !path_is_normalized(slink_name))
                 return log_device_debug_errno(dev, SYNTHETIC_ERRNO(EINVAL),
                                               "Invalid symbolic link of device node: %s", slink);
 
@@ -244,31 +283,31 @@ static int link_update(sd_device *dev, const char *slink, bool add) {
         if (r < 0)
                 return log_device_debug_errno(dev, r, "Failed to get device id: %m");
 
-        escape_path(slink_name, name_enc, sizeof(name_enc));
+        (void) udev_node_escape_path(slink_name, name_enc, sizeof(name_enc));
         dirname = path_join("/run/udev/links/", name_enc);
         if (!dirname)
-                return log_oom();
+                return log_oom_debug();
+
         filename = path_join(dirname, id);
         if (!filename)
-                return log_oom();
+                return log_oom_debug();
 
         if (!add) {
-                if (unlink(filename) == 0)
-                        (void) rmdir(dirname);
-        } else
-                for (;;) {
-                        _cleanup_close_ int fd = -1;
+                if (unlink(filename) < 0 && errno != ENOENT)
+                        log_device_debug_errno(dev, errno, "Failed to remove %s, ignoring: %m", filename);
 
-                        r = mkdir_parents(filename, 0755);
-                        if (!IN_SET(r, 0, -ENOENT))
-                                return r;
-
-                        fd = open(filename, O_WRONLY|O_CREAT|O_CLOEXEC|O_TRUNC|O_NOFOLLOW, 0444);
-                        if (fd >= 0)
+                (void) rmdir(dirname);
+        } else {
+                for (unsigned j = 0; j < TOUCH_FILE_MAX_RETRIES; j++) {
+                        /* This may fail with -ENOENT when the parent directory is removed during
+                         * creating the file by another udevd worker. */
+                        r = touch_file(filename, /* parents= */ true, USEC_INFINITY, UID_INVALID, GID_INVALID, 0444);
+                        if (r != -ENOENT)
                                 break;
-                        if (errno != ENOENT)
-                                return -errno;
                 }
+                if (r < 0)
+                        return log_device_debug_errno(dev, r, "Failed to create %s: %m", filename);
+        }
 
         /* If the database entry is not written yet we will just do one iteration and possibly wrong symlink
          * will be fixed in the second invocation. */
@@ -280,23 +319,25 @@ static int link_update(sd_device *dev, const char *slink, bool add) {
 
                 r = stat(dirname, &st1);
                 if (r < 0 && errno != ENOENT)
-                        return -errno;
+                        return log_device_debug_errno(dev, errno, "Failed to stat %s: %m", dirname);
 
                 r = link_find_prioritized(dev, add, dirname, &target);
-                if (r == -ENOENT) {
-                        log_device_debug(dev, "No reference left, removing '%s'", slink);
-                        if (unlink(slink) == 0)
-                                (void) rmdir_parents(slink, "/");
+                if (r < 0)
+                        return log_device_debug_errno(dev, r, "Failed to determine highest priority for symlink '%s': %m", slink);
+                if (r == 0) {
+                        log_device_debug(dev, "No reference left for '%s', removing", slink);
 
+                        if (unlink(slink) < 0 && errno != ENOENT)
+                                log_device_debug_errno(dev, errno, "Failed to remove '%s', ignoring: %m", slink);
+
+                        (void) rmdir_parents(slink, "/dev");
                         break;
-                } else if (r < 0)
-                        return log_device_error_errno(dev, r, "Failed to determine highest priority symlink: %m");
+                }
 
                 r = node_symlink(dev, target, slink);
-                if (r < 0) {
-                        (void) unlink(filename);
-                        break;
-                } else if (r == 1)
+                if (r < 0)
+                        return r;
+                if (r == 1)
                         /* We have replaced already existing symlink, possibly there is some other device trying
                          * to claim the same symlink. Let's do one more iteration to give us a chance to fix
                          * the error if other device actually claims the symlink with higher priority. */
@@ -306,7 +347,7 @@ static int link_update(sd_device *dev, const char *slink, bool add) {
                 if ((st1.st_mode & S_IFMT) != 0) {
                         r = stat(dirname, &st2);
                         if (r < 0 && errno != ENOENT)
-                                return -errno;
+                                return log_device_debug_errno(dev, errno, "Failed to stat %s: %m", dirname);
 
                         if (stat_inode_unmodified(&st1, &st2))
                                 break;
@@ -317,16 +358,12 @@ static int link_update(sd_device *dev, const char *slink, bool add) {
 }
 
 int udev_node_update_old_links(sd_device *dev, sd_device *dev_old) {
-        const char *name, *devpath;
+        const char *name;
         int r;
 
         assert(dev);
         assert(dev_old);
 
-        r = sd_device_get_devpath(dev, &devpath);
-        if (r < 0)
-                return log_device_debug_errno(dev, r, "Failed to get devpath: %m");
-
         /* update possible left-over symlinks */
         FOREACH_DEVICE_DEVLINK(dev_old, name) {
                 const char *name_current;
@@ -342,8 +379,10 @@ int udev_node_update_old_links(sd_device *dev, sd_device *dev_old) {
                 if (found)
                         continue;
 
-                log_device_debug(dev, "Updating old name, '%s' no longer belonging to '%s'",
-                                 name, devpath);
+                log_device_debug(dev,
+                                 "Updating old device symlink '%s', which is no longer belonging to this device.",
+                                 name);
+
                 r = link_update(dev, name, false);
                 if (r < 0)
                         log_device_warning_errno(dev, r,
@@ -477,7 +516,6 @@ static int node_permissions_apply(sd_device *dev, bool apply_mac,
 }
 
 static int xsprintf_dev_num_path_from_sd_device(sd_device *dev, char **ret) {
-        char filename[DEV_NUM_PATH_MAX], *s;
         const char *subsystem;
         dev_t devnum;
         int r;
@@ -492,16 +530,7 @@ static int xsprintf_dev_num_path_from_sd_device(sd_device *dev, char **ret) {
         if (r < 0)
                 return r;
 
-        xsprintf_dev_num_path(filename,
-                              streq(subsystem, "block") ? "block" : "char",
-                              devnum);
-
-        s = strdup(filename);
-        if (!s)
-                return -ENOMEM;
-
-        *ret = s;
-        return 0;
+        return device_path_make_major_minor(streq(subsystem, "block") ? S_IFBLK : S_IFCHR, devnum, ret);
 }
 
 int udev_node_add(sd_device *dev, bool apply,
@@ -528,13 +557,6 @@ int udev_node_add(sd_device *dev, bool apply,
         if (r < 0)
                 return r;
 
-        r = xsprintf_dev_num_path_from_sd_device(dev, &filename);
-        if (r < 0)
-                return log_device_debug_errno(dev, r, "Failed to get device path: %m");
-
-        /* always add /dev/{block,char}/$major:$minor */
-        (void) node_symlink(dev, devnode, filename);
-
         /* create/update symlinks, add symlinks to name index */
         FOREACH_DEVICE_DEVLINK(dev, devlink) {
                 r = link_update(dev, devlink, true);
@@ -544,6 +566,15 @@ int udev_node_add(sd_device *dev, bool apply,
                                                  devlink);
         }
 
+        r = xsprintf_dev_num_path_from_sd_device(dev, &filename);
+        if (r < 0)
+                return log_device_debug_errno(dev, r, "Failed to get device path: %m");
+
+        /* always add /dev/{block,char}/$major:$minor */
+        r = node_symlink(dev, devnode, filename);
+        if (r < 0)
+                return log_device_warning_errno(dev, r, "Failed to create device symlink '%s': %m", filename);
+
         return 0;
 }
 
@@ -568,7 +599,8 @@ int udev_node_remove(sd_device *dev) {
                 return log_device_debug_errno(dev, r, "Failed to get device path: %m");
 
         /* remove /dev/{block,char}/$major:$minor */
-        (void) unlink(filename);
+        if (unlink(filename) < 0 && errno != ENOENT)
+                return log_device_debug_errno(dev, errno, "Failed to remove '%s': %m", filename);
 
         return 0;
 }

src/udev/udev-node.h

@@ -13,3 +13,5 @@ int udev_node_add(sd_device *dev, bool apply,
                   OrderedHashmap *seclabel_list);
 int udev_node_remove(sd_device *dev);
 int udev_node_update_old_links(sd_device *dev, sd_device *dev_old);
+
+size_t udev_node_escape_path(const char *src, char *dest, size_t size);

tools/analyze-dump-sort.py

@@ -0,0 +1,78 @@
+#!/usr/bin/python
+# SPDX-License-Identifier: LGPL-2.1-or-later
+
+"""
+A helper to compare 'systemd-analyze dump' outputs.
+
+systemd-analyze dump >/var/tmp/dump1
+(reboot)
+tools/analyze-dump-sort.py /var/tmp/dump1 → this does a diff from dump1 to current
+
+systemd-analyze dump >/var/tmp/dump2
+tools/analyze-dump-sort.py /var/tmp/{dump1,dump2} → this does a diff from dump1 to dump2
+"""
+
+import argparse
+import tempfile
+import subprocess
+
+def sort_dump(sourcefile, destfile=None):
+    if destfile is None:
+        destfile = tempfile.NamedTemporaryFile('wt')
+
+    units = {}
+    unit = []
+
+    same = []
+
+    for line in sourcefile:
+        line = line.rstrip()
+
+        header = line.split(':')[0]
+        if 'Timestamp' in header or 'Invocation ID' in header or 'PID' in header:
+            line = header + ': …'
+
+        if line.startswith('->'):
+            if unit:
+                units[unit[0]] = unit
+            unit = [line]
+        elif line.startswith('\t'):
+            assert unit
+
+            if same and same[0].startswith(header):
+                same.append(line)
+            else:
+                unit.extend(sorted(same, key=str.lower))
+                same = [line]
+        else:
+            print(line, file=destfile)
+
+    if unit:
+        units[unit[0]] = unit
+
+    for unit in sorted(units.values()):
+        print('\n'.join(unit), file=destfile)
+
+    destfile.flush()
+    return destfile
+
+def parse_args():
+    p = argparse.ArgumentParser(description=__doc__)
+    p.add_argument('one')
+    p.add_argument('two', nargs='?')
+    p.add_argument('--user', action='store_true')
+    return p.parse_args()
+
+if __name__ == '__main__':
+    opts = parse_args()
+
+    one = sort_dump(open(opts.one))
+    if opts.two:
+        two = sort_dump(open(opts.two))
+    else:
+        user = ['--user'] if opts.user else []
+        two = subprocess.run(['systemd-analyze', 'dump', *user],
+                             capture_output=True, text=True, check=True)
+        two = sort_dump(two.stdout.splitlines())
+    with subprocess.Popen(['diff', '-U10', one.name, two.name], stdout=subprocess.PIPE) as diff:
+        subprocess.Popen(['less'], stdin=diff.stdout)