diff options
56 files changed, 1148 insertions, 837 deletions
diff --git a/DISTRO_PORTING b/DISTRO_PORTING index a397d700c6..9f61bd6de6 100644 --- a/DISTRO_PORTING +++ b/DISTRO_PORTING @@ -41,6 +41,13 @@ NTP POOL: NTP servers, then you will get served wrong time, and will rely on services that might not be supported for long. +PAM: + The default PAM config shipped by systemd is really bare bones. + It does not include many modules your distro might want to enable + to provide a more seamless experience. For example, limits set in + /etc/security/limits.conf will not be read unless you load pam_limits. + Make sure you add modules your distro expects from user services. + CONTRIBUTING UPSTREAM: We generally do no longer accept distribution-specific diff --git a/Makefile.am b/Makefile.am index f7ae9ff7aa..b09b0cf167 100644 --- a/Makefile.am +++ b/Makefile.am @@ -524,7 +524,6 @@ dist_systemunit_DATA_busnames += \ nodist_systemunit_DATA = \ units/getty@.service \ units/serial-getty@.service \ - units/console-shell.service \ units/console-getty.service \ units/container-getty@.service \ units/systemd-initctl.service \ @@ -587,7 +586,6 @@ dist_systempreset_DATA = \ EXTRA_DIST += \ units/getty@.service.m4 \ units/serial-getty@.service.m4 \ - units/console-shell.service.m4.in \ units/console-getty.service.m4.in \ units/container-getty@.service.m4.in \ units/rescue.service.in \ @@ -32,6 +32,8 @@ Janitorial Clean-ups: Features: +* on cgroupsv2 add DelegateControllers=, to pick the precise cgroup controllers to delegate + * in networkd, when matching device types, fix up DEVTYPE rubbish the kernel passes to us * enable LockMLOCK to take a percentage value relative to physical memory diff --git a/man/resolved.conf.xml b/man/resolved.conf.xml index a7710dcb44..4fc1ef1b33 100644 --- a/man/resolved.conf.xml +++ b/man/resolved.conf.xml @@ -213,6 +213,18 @@ (such as 127.0.0.1 or ::1), in order to avoid duplicate local caching.</para></listitem> </varlistentry> + <varlistentry> + <term><varname>DNSStubListener=</varname></term> + <listitem><para>Takes a boolean argument or one of <literal>udp</literal> and <literal>tcp</literal>. If + <literal>udp</literal> (the default), a DNS stub resolver will listen for UDP requests on address 127.0.0.53 + port 53. If <literal>tcp</literal>, the stub will listen for TCP requests on the same address and port. If + <literal>yes</literal>, the stub listens for both UDP and TCP requests. If <literal>no</literal>, the stub + listener is disabled.</para> + + <para>Note that the DNS stub listener is turned off implicitly when its listening address and port are already + in use.</para></listitem> + </varlistentry> + </variablelist> </refsect1> diff --git a/man/systemd-journal-remote.xml b/man/systemd-journal-remote.xml index f208f8deb4..ee2d5c2486 100644 --- a/man/systemd-journal-remote.xml +++ b/man/systemd-journal-remote.xml @@ -250,20 +250,19 @@ </varlistentry> <varlistentry> - <term><option>--compress</option></term> - <term><option>--no-compress</option></term> + <term><option>--compress</option> [<replaceable>BOOL</replaceable>]</term> - <listitem><para>Compress or not, respectively, the data in the - journal using XZ.</para></listitem> + <listitem><para>If this is set to <literal>yes</literal> then compress + the data in the journal using XZ. The default is <literal>yes</literal>. + </para></listitem> </varlistentry> <varlistentry> - <term><option>--seal</option></term> - <term><option>--no-seal</option></term> + <term><option>--seal</option> [<replaceable>BOOL</replaceable>]</term> - <listitem><para>Periodically sign or not, respectively, the - data in the journal using Forward Secure Sealing. - </para></listitem> + <listitem><para>If this is set to <literal>yes</literal> then + periodically sign the data in the journal using Forward Secure Sealing. + The default is <literal>no</literal>.</para></listitem> </varlistentry> <varlistentry> diff --git a/man/systemd-nspawn.xml b/man/systemd-nspawn.xml index bf3860604c..5ac54df81a 100644 --- a/man/systemd-nspawn.xml +++ b/man/systemd-nspawn.xml @@ -405,35 +405,36 @@ purposes (usually in the range beyond the host's UID/GID 65536). The parameter may be specified as follows:</para> <orderedlist> - <listitem><para>The value <literal>no</literal> turns off user namespacing. This is the default.</para></listitem> - - <listitem><para>The value <literal>yes</literal> (or the omission of a parameter) turns on user - namespacing. The UID/GID range to use is determined automatically from the file ownership of the root - directory of the container's directory tree. To use this option, make sure to prepare the directory tree in - advance, and ensure that all files and directories in it are owned by UIDs/GIDs in the range you'd like to - use. Also, make sure that used file ACLs exclusively reference UIDs/GIDs in the appropriate range. If this - mode is used the number of UIDs/GIDs assigned to the container for use is 65536, and the UID/GID of the - root directory must be a multiple of 65536.</para></listitem> - - <listitem><para>The value "pick" turns on user namespacing. In this case the UID/GID range is automatically - chosen. As first step, the file owner of the root directory of the container's directory tree is read, and it - is checked that it is currently not used by the system otherwise (in particular, that no other container is - using it). If this check is successful, the UID/GID range determined this way is used, similar to the - behaviour if "yes" is specified. If the check is not successful (and thus the UID/GID range indicated in the - root directory's file owner is already used elsewhere) a new – currently unused – UID/GID range of 65536 - UIDs/GIDs is randomly chosen between the host UID/GIDs of 524288 and 1878982656, always starting at a - multiple of 65536. This setting implies <option>--private-users-chown</option> (see below), which has the - effect that the files and directories in the container's directory tree will be owned by the appropriate - users of the range picked. Using this option makes user namespace behaviour fully automatic. Note that the - first invocation of a previously unused container image might result in picking a new UID/GID range for it, - and thus in the (possibly expensive) file ownership adjustment operation. However, subsequent invocations of - the container will be cheap (unless of course the picked UID/GID range is assigned to a different use by - then).</para></listitem> - - <listitem><para>Finally if one or two colon-separated numeric parameters are specified, user namespacing is - turned on, too. The first parameter specifies the first host UID/GID to assign to the container, the second - parameter specifies the number of host UIDs/GIDs to assign to the container. If the second parameter is - omitted, 65536 UIDs/GIDs are assigned.</para></listitem> + <listitem><para>If one or two colon-separated numers are specified, user namespacing is turned on. The first + parameter specifies the first host UID/GID to assign to the container, the second parameter specifies the + number of host UIDs/GIDs to assign to the container. If the second parameter is omitted, 65536 UIDs/GIDs are + assigned.</para></listitem> + + <listitem><para>If the parameter is omitted, or true, user namespacing is turned on. The UID/GID range to + use is determined automatically from the file ownership of the root directory of the container's directory + tree. To use this option, make sure to prepare the directory tree in advance, and ensure that all files and + directories in it are owned by UIDs/GIDs in the range you'd like to use. Also, make sure that used file ACLs + exclusively reference UIDs/GIDs in the appropriate range. If this mode is used the number of UIDs/GIDs + assigned to the container for use is 65536, and the UID/GID of the root directory must be a multiple of + 65536.</para></listitem> + + <listitem><para>If the parameter is false, user namespacing is turned off. This is the default.</para> + </listitem> + + <listitem><para>The special value <literal>pick</literal> turns on user namespacing. In this case the UID/GID + range is automatically chosen. As first step, the file owner of the root directory of the container's + directory tree is read, and it is checked that it is currently not used by the system otherwise (in + particular, that no other container is using it). If this check is successful, the UID/GID range determined + this way is used, similar to the behaviour if "yes" is specified. If the check is not successful (and thus + the UID/GID range indicated in the root directory's file owner is already used elsewhere) a new – currently + unused – UID/GID range of 65536 UIDs/GIDs is randomly chosen between the host UID/GIDs of 524288 and + 1878982656, always starting at a multiple of 65536. This setting implies + <option>--private-users-chown</option> (see below), which has the effect that the files and directories in + the container's directory tree will be owned by the appropriate users of the range picked. Using this option + makes user namespace behaviour fully automatic. Note that the first invocation of a previously unused + container image might result in picking a new UID/GID range for it, and thus in the (possibly expensive) file + ownership adjustment operation. However, subsequent invocations of the container will be cheap (unless of + course the picked UID/GID range is assigned to a different use by then).</para></listitem> </orderedlist> <para>It is recommended to assign at least 65536 UIDs/GIDs to each container, so that the usable UID/GID range in the @@ -454,17 +455,6 @@ </varlistentry> <varlistentry> - <term><option>-U</option></term> - - <listitem><para>If the kernel supports the user namespaces feature, equivalent to - <option>--private-users=pick</option>, otherwise equivalent to - <option>--private-users=no</option>.</para> - - <para>Note that <option>-U</option> is the default if the <filename>systemd-nspawn@.service</filename> template unit - file is used.</para></listitem> - </varlistentry> - - <varlistentry> <term><option>--private-users-chown</option></term> <listitem><para>If specified, all files and directories in the container's directory tree will adjusted so that @@ -477,6 +467,23 @@ </varlistentry> <varlistentry> + <term><option>-U</option></term> + + <listitem><para>If the kernel supports the user namespaces feature, equivalent to + <option>--private-users=pick --private-users-chown</option>, otherwise equivalent to + <option>--private-users=no</option>.</para> + + <para>Note that <option>-U</option> is the default if the + <filename>systemd-nspawn@.service</filename> template unit file is used.</para> + + <para>Note: it is possible to undo the effect of <option>--private-users-chown</option> (or + <option>-U</option>) on the file system by redoing the operation with the first UID of 0:</para> + + <programlisting>systemd-nspawn … --private-users=0 --private-users-chown</programlisting> + </listitem> + </varlistentry> + + <varlistentry> <term><option>--private-network</option></term> <listitem><para>Disconnect networking of the container from diff --git a/man/systemd.netdev.xml b/man/systemd.netdev.xml index e378e61dd1..ffb66e735b 100644 --- a/man/systemd.netdev.xml +++ b/man/systemd.netdev.xml @@ -537,7 +537,7 @@ </listitem> </varlistentry> <varlistentry> - <term><varname>UDPCheckSum=</varname></term> + <term><varname>UDPChecksum=</varname></term> <listitem> <para>A boolean. When true, transmitting UDP checksums when doing VXLAN/IPv4 is turned on.</para> </listitem> @@ -549,19 +549,19 @@ </listitem> </varlistentry> <varlistentry> - <term><varname>UDP6ZeroCheckSumRx=</varname></term> + <term><varname>UDP6ZeroChecksumRx=</varname></term> <listitem> <para>A boolean. When true, receiving zero checksums in VXLAN/IPv6 is turned on.</para> </listitem> </varlistentry> <varlistentry> - <term><varname>RemoteCheckSumTx=</varname></term> + <term><varname>RemoteChecksumTx=</varname></term> <listitem> <para>A boolean. When true, remote transmit checksum offload of VXLAN is turned on.</para> </listitem> </varlistentry> <varlistentry> - <term><varname>RemoteCheckSumRx=</varname></term> + <term><varname>RemoteChecksumRx=</varname></term> <listitem> <para>A boolean. When true, remote receive checksum offload in VXLAN is turned on.</para> </listitem> diff --git a/man/systemd.network.xml b/man/systemd.network.xml index 0af927db19..2fb4907634 100644 --- a/man/systemd.network.xml +++ b/man/systemd.network.xml @@ -695,6 +695,57 @@ which is then configured to use them explicitly.</para> </listitem> </varlistentry> + <varlistentry> + <term><varname>HomeAddress=</varname></term> + <listitem> + <para>Takes a boolean argument. Designates this address the "home address" as defined in + <ulink url="https://tools.ietf.org/html/rfc6275">RFC 6275</ulink>. + Supported only on IPv6. Defaults to false.</para> + </listitem> + </varlistentry> + <varlistentry> + <term><varname>DuplicateAddressDetection=</varname></term> + <listitem> + <para>Takes a boolean argument. Do not perform Duplicate Address Detection + <ulink url="https://tools.ietf.org/html/rfc4862">RFC 4862</ulink> when adding this address. + Supported only on IPv6. Defaults to false.</para> + </listitem> + </varlistentry> + <varlistentry> + <term><varname>ManageTemporaryAddress=</varname></term> + <listitem> + <para>Takes a boolean argument. If true the kernel manage temporary addresses created + from this one as template on behalf of Privacy Extensions + <ulink url="https://tools.ietf.org/html/rfc3041">RFC 3041</ulink>. For this to become + active, the use_tempaddr sysctl setting has to be set to a value greater than zero. + The given address needs to have a prefix length of 64. This flag allows to use privacy + extensions in a manually configured network, just like if stateless auto-configuration + was active. Defaults to false. </para> + </listitem> + </varlistentry> + <varlistentry> + <term><varname>PrefixRoute=</varname></term> + <listitem> + <para>Takes a boolean argument. When adding or modifying an IPv6 address, the userspace + application needs a way to suppress adding a prefix route. This is for example relevant + together with IFA_F_MANAGERTEMPADDR, where userspace creates autoconf generated addresses, + but depending on on-link, no route for the prefix should be added. Defaults to false.</para> + </listitem> + </varlistentry> + <varlistentry> + <term><varname>AutoJoin=</varname></term> + <listitem> + <para>Takes a boolean argument. Joining multicast group on ethernet level via + <command>ip maddr</command> command would not work if we have an Ethernet switch that does + IGMP snooping since the switch would not replicate multicast packets on ports that did not + have IGMP reports for the multicast addresses. Linux vxlan interfaces created via + <command>ip link add vxlan</command> or networkd's netdev kind vxlan have the group option + that enables then to do the required join. By extending ip address command with option + <literal>autojoin</literal> we can get similar functionality for openvswitch (OVS) vxlan + interfaces as well as other tunneling mechanisms that need to receive multicast traffic. + Defaults to <literal>no</literal>.</para> + </listitem> + </varlistentry> </variablelist> </refsect1> diff --git a/man/systemd.xml b/man/systemd.xml index e30333e209..7f24a874ed 100644 --- a/man/systemd.xml +++ b/man/systemd.xml @@ -272,7 +272,7 @@ <title>Concepts</title> <para>systemd provides a dependency system between various - entities called "units" of 12 different types. Units encapsulate + entities called "units" of 11 different types. Units encapsulate various objects that are relevant for system boot-up and maintenance. The majority of units are configured in unit configuration files, whose syntax and basic set of options is diff --git a/shell-completion/bash/systemctl.in b/shell-completion/bash/systemctl.in index 2a45dcbba0..dcf71a1f51 100644 --- a/shell-completion/bash/systemctl.in +++ b/shell-completion/bash/systemctl.in @@ -41,7 +41,7 @@ __contains_word () { __filter_units_by_property () { local mode=$1 property=$2 value=$3 ; shift 3 local units=("$@") - local props + local props i IFS=$'\n' read -rd '' -a props < \ <(__systemctl $mode show --property "$property" -- "${units[@]}") for ((i=0; $i < ${#units[*]}; i++)); do @@ -51,6 +51,33 @@ __filter_units_by_property () { done } +__filter_units_by_properties () { + local mode=$1 properties=$2 values=$3 ; shift 3 + local units=("$@") + local props i j conditions=() + IFS=$'\n' read -rd '' -a props < \ + <(__systemctl $mode show --property "$properties" -- "${units[@]}") + IFS=$',' read -r -a properties < <(echo $properties) + IFS=$',' read -r -a values < <(echo $values) + for ((i=0; i < ${#properties[*]}; i++)); do + for ((j=0; j < ${#properties[*]}; j++)); do + if [[ ${props[i]%%=*} == ${properties[j]} ]]; then + conditions+=( "${properties[j]}=${values[j]}" ) + fi + done + done + for ((i=0; i < ${#units[*]}; i++)); do + for ((j=0; j < ${#conditions[*]}; j++)); do + if [[ "${props[ i * ${#conditions[*]} + j]}" != "${conditions[j]}" ]]; then + break + fi + done + if (( j == ${#conditions[*]} )); then + echo " ${units[i]}" + fi + done +} + __get_all_units () { { __systemctl $1 list-unit-files; __systemctl $1 list-units --all; } \ | { while read -r a b; do [[ $a =~ @\. ]] || echo " $a"; done; }; } __get_template_names () { __systemctl $1 list-unit-files \ @@ -60,12 +87,12 @@ __get_active_units () { __systemctl $1 list-units \ | { while read -r a b; do echo " $a"; done; }; } __get_startable_units () { # find startable inactive units - __filter_units_by_property $mode ActiveState inactive $( - __filter_units_by_property $mode CanStart yes $( - __systemctl $mode list-unit-files --state enabled,disabled,static | \ - { while read -r a b; do [[ $a =~ @\. ]] || echo " $a"; done; } - __systemctl $mode list-units --state inactive,failed | \ - { while read -r a b; do echo " $a"; done; } )) + __filter_units_by_properties $mode ActiveState,CanStart inactive,yes $( + { __systemctl $mode list-unit-files --state enabled,enabled-runtime,linked,linked-runtime,static,indirect,disabled,generated,transient | \ + { while read -r a b; do [[ $a =~ @\. ]] || echo " $a"; done; } + __systemctl $mode list-units --state inactive,failed | \ + { while read -r a b c; do [[ $b == "loaded" ]] && echo " $a"; done; } + } | sort -u ) } __get_restartable_units () { # filter out masked and not-found diff --git a/src/basic/exit-status.c b/src/basic/exit-status.c index d488cfc59f..59557f8afe 100644 --- a/src/basic/exit-status.c +++ b/src/basic/exit-status.c @@ -24,12 +24,12 @@ #include "macro.h" #include "set.h" -const char* exit_status_to_string(ExitStatus status, ExitStatusLevel level) { +const char* exit_status_to_string(int status, ExitStatusLevel level) { /* We cast to int here, so that -Wenum doesn't complain that * EXIT_SUCCESS/EXIT_FAILURE aren't in the enum */ - switch ((int) status) { + switch (status) { case EXIT_SUCCESS: return "SUCCESS"; @@ -39,7 +39,7 @@ const char* exit_status_to_string(ExitStatus status, ExitStatusLevel level) { } if (IN_SET(level, EXIT_STATUS_SYSTEMD, EXIT_STATUS_LSB)) { - switch ((int) status) { + switch (status) { case EXIT_CHDIR: return "CHDIR"; @@ -140,19 +140,19 @@ const char* exit_status_to_string(ExitStatus status, ExitStatusLevel level) { case EXIT_RUNTIME_DIRECTORY: return "RUNTIME_DIRECTORY"; - case EXIT_CHOWN: - return "CHOWN"; - case EXIT_MAKE_STARTER: return "MAKE_STARTER"; + case EXIT_CHOWN: + return "CHOWN"; + case EXIT_SMACK_PROCESS_LABEL: return "SMACK_PROCESS_LABEL"; } } if (level == EXIT_STATUS_LSB) { - switch ((int) status) { + switch (status) { case EXIT_INVALIDARGUMENT: return "INVALIDARGUMENT"; @@ -177,34 +177,23 @@ const char* exit_status_to_string(ExitStatus status, ExitStatusLevel level) { return NULL; } - -bool is_clean_exit(int code, int status, ExitStatusSet *success_status) { +bool is_clean_exit(int code, int status, ExitClean clean, ExitStatusSet *success_status) { if (code == CLD_EXITED) return status == 0 || (success_status && set_contains(success_status->status, INT_TO_PTR(status))); - /* If a daemon does not implement handlers for some of the - * signals that's not considered an unclean shutdown */ + /* If a daemon does not implement handlers for some of the signals that's not considered an unclean shutdown */ if (code == CLD_KILLED) - return IN_SET(status, SIGHUP, SIGINT, SIGTERM, SIGPIPE) || + return + (clean == EXIT_CLEAN_DAEMON && IN_SET(status, SIGHUP, SIGINT, SIGTERM, SIGPIPE)) || (success_status && set_contains(success_status->signal, INT_TO_PTR(status))); return false; } -bool is_clean_exit_lsb(int code, int status, ExitStatusSet *success_status) { - - if (is_clean_exit(code, status, success_status)) - return true; - - return - code == CLD_EXITED && - IN_SET(status, EXIT_NOTINSTALLED, EXIT_NOTCONFIGURED); -} - void exit_status_set_free(ExitStatusSet *x) { assert(x); diff --git a/src/basic/exit-status.h b/src/basic/exit-status.h index 2309f68815..0cfdfd7891 100644 --- a/src/basic/exit-status.h +++ b/src/basic/exit-status.h @@ -31,7 +31,7 @@ * https://refspecs.linuxbase.org/LSB_5.0.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html */ -typedef enum ExitStatus { +enum { /* EXIT_SUCCESS defined by libc */ /* EXIT_FAILURE defined by libc */ EXIT_INVALIDARGUMENT = 2, @@ -82,7 +82,7 @@ typedef enum ExitStatus { EXIT_MAKE_STARTER, EXIT_CHOWN, EXIT_SMACK_PROCESS_LABEL, -} ExitStatus; +}; typedef enum ExitStatusLevel { EXIT_STATUS_MINIMAL, /* only cover libc EXIT_STATUS/EXIT_FAILURE */ @@ -96,10 +96,14 @@ typedef struct ExitStatusSet { Set *signal; } ExitStatusSet; -const char* exit_status_to_string(ExitStatus status, ExitStatusLevel level) _const_; +const char* exit_status_to_string(int status, ExitStatusLevel level) _const_; -bool is_clean_exit(int code, int status, ExitStatusSet *success_status); -bool is_clean_exit_lsb(int code, int status, ExitStatusSet *success_status); +typedef enum ExitClean { + EXIT_CLEAN_DAEMON, + EXIT_CLEAN_COMMAND, +} ExitClean; + +bool is_clean_exit(int code, int status, ExitClean clean, ExitStatusSet *success_status); void exit_status_set_free(ExitStatusSet *x); bool exit_status_set_is_empty(ExitStatusSet *x); diff --git a/src/basic/missing.h b/src/basic/missing.h index 13ff51cd35..4a78269e33 100644 --- a/src/basic/missing.h +++ b/src/basic/missing.h @@ -1052,6 +1052,10 @@ typedef int32_t key_serial_t; #define ETHERTYPE_LLDP 0x88cc #endif +#ifndef IFA_F_MCAUTOJOIN +#define IFA_F_MCAUTOJOIN 0x400 +#endif + #endif #include "missing_syscall.h" diff --git a/src/basic/path-util.c b/src/basic/path-util.c index b2fa81a294..c32e961af4 100644 --- a/src/basic/path-util.c +++ b/src/basic/path-util.c @@ -34,9 +34,11 @@ #include "alloc-util.h" #include "extract-word.h" #include "fs-util.h" +#include "glob-util.h" #include "log.h" #include "macro.h" #include "missing.h" +#include "parse-util.h" #include "path-util.h" #include "stat-util.h" #include "string-util.h" @@ -814,3 +816,69 @@ bool is_device_path(const char *path) { path_startswith(path, "/dev/") || path_startswith(path, "/sys/"); } + +int systemd_installation_has_version(const char *root, unsigned minimal_version) { + const char *pattern; + int r; + + /* Try to guess if systemd installation is later than the specified version. This + * is hacky and likely to yield false negatives, particularly if the installation + * is non-standard. False positives should be relatively rare. + */ + + NULSTR_FOREACH(pattern, + /* /lib works for systems without usr-merge, and for systems with a sane + * usr-merge, where /lib is a symlink to /usr/lib. /usr/lib is necessary + * for Gentoo which does a merge without making /lib a symlink. + */ + "lib/systemd/libsystemd-shared-*.so\0" + "usr/lib/systemd/libsystemd-shared-*.so\0") { + + _cleanup_strv_free_ char **names = NULL; + _cleanup_free_ char *path = NULL; + char *c, **name; + + path = prefix_root(root, pattern); + if (!path) + return -ENOMEM; + + r = glob_extend(&names, path); + if (r == -ENOENT) + continue; + if (r < 0) + return r; + + assert_se((c = endswith(path, "*.so"))); + *c = '\0'; /* truncate the glob part */ + + STRV_FOREACH(name, names) { + /* This is most likely to run only once, hence let's not optimize anything. */ + char *t, *t2; + unsigned version; + + t = startswith(*name, path); + if (!t) + continue; + + t2 = endswith(t, ".so"); + if (!t2) + continue; + + t2[0] = '\0'; /* truncate the suffix */ + + r = safe_atou(t, &version); + if (r < 0) { + log_debug_errno(r, "Found libsystemd shared at \"%s.so\", but failed to parse version: %m", *name); + continue; + } + + log_debug("Found libsystemd shared at \"%s.so\", version %u (%s).", + *name, version, + version >= minimal_version ? "OK" : "too old"); + if (version >= minimal_version) + return true; + } + } + + return false; +} diff --git a/src/basic/path-util.h b/src/basic/path-util.h index a27c13fcc3..78472f0961 100644 --- a/src/basic/path-util.h +++ b/src/basic/path-util.h @@ -125,3 +125,5 @@ char *file_in_same_dir(const char *path, const char *filename); bool hidden_or_backup_file(const char *filename) _pure_; bool is_device_path(const char *path); + +int systemd_installation_has_version(const char *root, unsigned minimal_version); diff --git a/src/boot/efi/measure.c b/src/boot/efi/measure.c index 7c016387c1..4ac11a9bb0 100644 --- a/src/boot/efi/measure.c +++ b/src/boot/efi/measure.c @@ -209,12 +209,35 @@ static EFI_STATUS tpm1_measure_to_pcr_and_event_log(const EFI_TCG *tcg, UINT32 p return EFI_SUCCESS; } +/* + * According to TCG EFI Protocol Specification for TPM 2.0 family, + * all events generated after the invocation of EFI_TCG2_GET_EVENT_LOG + * shall be stored in an instance of an EFI_CONFIGURATION_TABLE aka + * EFI TCG 2.0 final events table. Hence, it is necessary to trigger the + * internal switch through calling get_event_log() in order to allow + * to retrieve the logs from OS runtime. + */ +static EFI_STATUS trigger_tcg2_final_events_table(const EFI_TCG2 *tcg) +{ + return uefi_call_wrapper(tcg->GetEventLog, 5, tcg, + EFI_TCG2_EVENT_LOG_FORMAT_TCG_2, NULL, + NULL, NULL); +} static EFI_STATUS tpm2_measure_to_pcr_and_event_log(const EFI_TCG2 *tcg, UINT32 pcrindex, const EFI_PHYSICAL_ADDRESS buffer, UINT64 buffer_size, const CHAR16 *description) { EFI_STATUS status; EFI_TCG2_EVENT *tcg_event; UINTN desc_len; + static BOOLEAN triggered = FALSE; + + if (triggered == FALSE) { + status = trigger_tcg2_final_events_table(tcg); + if (EFI_ERROR(status)) + return status; + + triggered = TRUE; + } desc_len = StrLen(description) * sizeof(CHAR16); diff --git a/src/core/busname.c b/src/core/busname.c index 63c7dde0bd..b96ec09e67 100644 --- a/src/core/busname.c +++ b/src/core/busname.c @@ -872,7 +872,7 @@ static void busname_sigchld_event(Unit *u, pid_t pid, int code, int status) { n->control_pid = 0; - if (is_clean_exit(code, status, NULL)) + if (is_clean_exit(code, status, EXIT_CLEAN_COMMAND, NULL)) f = BUSNAME_SUCCESS; else if (code == CLD_EXITED) f = BUSNAME_FAILURE_EXIT_CODE; diff --git a/src/core/main.c b/src/core/main.c index 6fe440277e..4b82a57b3c 100644 --- a/src/core/main.c +++ b/src/core/main.c @@ -203,7 +203,7 @@ noreturn static void crash(int sig) { pid, sigchld_code_to_string(status.si_code), status.si_status, strna(status.si_code == CLD_EXITED - ? exit_status_to_string(status.si_status, EXIT_STATUS_FULL) + ? exit_status_to_string(status.si_status, EXIT_STATUS_MINIMAL) : signal_to_string(status.si_status))); else log_emergency("Caught <%s>, dumped core as pid "PID_FMT".", signal_to_string(sig), pid); @@ -1532,15 +1532,9 @@ int main(int argc, char *argv[]) { * need to do that for user instances since they never log * into the console. */ log_show_color(colors_enabled()); - make_null_stdio(); - } - - /* Initialize default unit */ - r = free_and_strdup(&arg_default_unit, SPECIAL_DEFAULT_TARGET); - if (r < 0) { - log_emergency_errno(r, "Failed to set default unit %s: %m", SPECIAL_DEFAULT_TARGET); - error_message = "Failed to set default unit"; - goto finish; + r = make_null_stdio(); + if (r < 0) + log_warning_errno(r, "Failed to redirect standard streams to /dev/null: %m"); } r = initialize_join_controllers(); @@ -1590,6 +1584,16 @@ int main(int argc, char *argv[]) { goto finish; } + /* Initialize default unit */ + if (!arg_default_unit) { + arg_default_unit = strdup(SPECIAL_DEFAULT_TARGET); + if (!arg_default_unit) { + r = log_oom(); + error_message = "Failed to set default unit"; + goto finish; + } + } + if (arg_action == ACTION_TEST && geteuid() == 0) { log_error("Don't run test mode as root."); diff --git a/src/core/mount.c b/src/core/mount.c index 436c0e1029..15619dffe3 100644 --- a/src/core/mount.c +++ b/src/core/mount.c @@ -830,7 +830,7 @@ static void mount_enter_signal(Mount *m, MountState state, MountResult f) { fail: log_unit_warning_errno(UNIT(m), r, "Failed to kill processes: %m"); - if (state == MOUNT_REMOUNTING_SIGTERM || state == MOUNT_REMOUNTING_SIGKILL) + if (IN_SET(state, MOUNT_REMOUNTING_SIGTERM, MOUNT_REMOUNTING_SIGKILL)) mount_enter_mounted(m, MOUNT_FAILURE_RESOURCES); else mount_enter_dead(m, MOUNT_FAILURE_RESOURCES); @@ -986,18 +986,19 @@ static int mount_start(Unit *u) { /* We cannot fulfill this request right now, try again later * please! */ - if (m->state == MOUNT_UNMOUNTING || - m->state == MOUNT_UNMOUNTING_SIGTERM || - m->state == MOUNT_UNMOUNTING_SIGKILL || - m->state == MOUNT_MOUNTING_SIGTERM || - m->state == MOUNT_MOUNTING_SIGKILL) + if (IN_SET(m->state, + MOUNT_UNMOUNTING, + MOUNT_UNMOUNTING_SIGTERM, + MOUNT_UNMOUNTING_SIGKILL, + MOUNT_MOUNTING_SIGTERM, + MOUNT_MOUNTING_SIGKILL)) return -EAGAIN; /* Already on it! */ if (m->state == MOUNT_MOUNTING) return 0; - assert(m->state == MOUNT_DEAD || m->state == MOUNT_FAILED); + assert(IN_SET(m->state, MOUNT_DEAD, MOUNT_FAILED)); r = unit_start_limit_test(u); if (r < 0) { @@ -1023,19 +1024,21 @@ static int mount_stop(Unit *u) { assert(m); /* Already on it */ - if (m->state == MOUNT_UNMOUNTING || - m->state == MOUNT_UNMOUNTING_SIGKILL || - m->state == MOUNT_UNMOUNTING_SIGTERM || - m->state == MOUNT_MOUNTING_SIGTERM || - m->state == MOUNT_MOUNTING_SIGKILL) + if (IN_SET(m->state, + MOUNT_UNMOUNTING, + MOUNT_UNMOUNTING_SIGKILL, + MOUNT_UNMOUNTING_SIGTERM, + MOUNT_MOUNTING_SIGTERM, + MOUNT_MOUNTING_SIGKILL)) return 0; - assert(m->state == MOUNT_MOUNTING || - m->state == MOUNT_MOUNTING_DONE || - m->state == MOUNT_MOUNTED || - m->state == MOUNT_REMOUNTING || - m->state == MOUNT_REMOUNTING_SIGTERM || - m->state == MOUNT_REMOUNTING_SIGKILL); + assert(IN_SET(m->state, + MOUNT_MOUNTING, + MOUNT_MOUNTING_DONE, + MOUNT_MOUNTED, + MOUNT_REMOUNTING, + MOUNT_REMOUNTING_SIGTERM, + MOUNT_REMOUNTING_SIGKILL)); mount_enter_unmounting(m); return 1; @@ -1163,7 +1166,7 @@ static void mount_sigchld_event(Unit *u, pid_t pid, int code, int status) { m->control_pid = 0; - if (is_clean_exit(code, status, NULL)) + if (is_clean_exit(code, status, EXIT_CLEAN_COMMAND, NULL)) f = MOUNT_SUCCESS; else if (code == CLD_EXITED) f = MOUNT_FAILURE_EXIT_CODE; @@ -1201,9 +1204,10 @@ static void mount_sigchld_event(Unit *u, pid_t pid, int code, int status) { case MOUNT_MOUNTING_SIGKILL: case MOUNT_MOUNTING_SIGTERM: - if (f == MOUNT_SUCCESS) - mount_enter_mounted(m, f); - else if (m->from_proc_self_mountinfo) + if (f == MOUNT_SUCCESS || m->from_proc_self_mountinfo) + /* If /bin/mount returned success, or if we see the mount point in /proc/self/mountinfo we are + * happy. If we see the first condition first, we should see the the second condition + * immediately after – or /bin/mount lies to us and is broken. */ mount_enter_mounted(m, f); else mount_enter_dead(m, f); diff --git a/src/core/service.c b/src/core/service.c index 8ce25c494c..63045ede55 100644 --- a/src/core/service.c +++ b/src/core/service.c @@ -2604,8 +2604,7 @@ static void service_sigchld_event(Unit *u, pid_t pid, int code, int status) { assert(s); assert(pid >= 0); - if (UNIT(s)->fragment_path ? is_clean_exit(code, status, &s->success_status) : - is_clean_exit_lsb(code, status, &s->success_status)) + if (is_clean_exit(code, status, s->type == SERVICE_ONESHOT ? EXIT_CLEAN_COMMAND : EXIT_CLEAN_DAEMON, &s->success_status)) f = SERVICE_SUCCESS; else if (code == CLD_EXITED) f = SERVICE_FAILURE_EXIT_CODE; diff --git a/src/core/socket.c b/src/core/socket.c index ae8a1f751f..0b1c4acfec 100644 --- a/src/core/socket.c +++ b/src/core/socket.c @@ -2746,7 +2746,7 @@ static void socket_sigchld_event(Unit *u, pid_t pid, int code, int status) { s->control_pid = 0; - if (is_clean_exit(code, status, NULL)) + if (is_clean_exit(code, status, EXIT_CLEAN_COMMAND, NULL)) f = SOCKET_SUCCESS; else if (code == CLD_EXITED) f = SOCKET_FAILURE_EXIT_CODE; diff --git a/src/core/swap.c b/src/core/swap.c index 0333eaefb8..b592abb9fb 100644 --- a/src/core/swap.c +++ b/src/core/swap.c @@ -992,7 +992,7 @@ static void swap_sigchld_event(Unit *u, pid_t pid, int code, int status) { s->control_pid = 0; - if (is_clean_exit(code, status, NULL)) + if (is_clean_exit(code, status, EXIT_CLEAN_COMMAND, NULL)) f = SWAP_SUCCESS; else if (code == CLD_EXITED) f = SWAP_FAILURE_EXIT_CODE; diff --git a/src/journal-remote/journal-remote.c b/src/journal-remote/journal-remote.c index a9009cfefe..d86c3681b1 100644 --- a/src/journal-remote/journal-remote.c +++ b/src/journal-remote/journal-remote.c @@ -127,6 +127,10 @@ static int spawn_child(const char* child, char** argv) { if (r < 0) log_warning_errno(errno, "Failed to close write end of pipe: %m"); + r = fd_nonblock(fd[0], true); + if (r < 0) + log_warning_errno(errno, "Failed to set child pipe to non-blocking: %m"); + return fd[0]; } diff --git a/src/login/systemd-user.m4 b/src/login/systemd-user.m4 index f188a8e548..fe38b24fef 100644 --- a/src/login/systemd-user.m4 +++ b/src/login/systemd-user.m4 @@ -2,11 +2,9 @@ # # Used by systemd --user instances. -account include system-auth - m4_ifdef(`HAVE_SELINUX', session required pam_selinux.so close session required pam_selinux.so nottys open )m4_dnl session required pam_loginuid.so -session include system-auth +session optional pam_systemd.so diff --git a/src/machine/machinectl.c b/src/machine/machinectl.c index e9de31e184..7b9be3b425 100644 --- a/src/machine/machinectl.c +++ b/src/machine/machinectl.c @@ -2629,6 +2629,7 @@ static int clean_images(int argc, char *argv[], void *userdata) { } static int help(int argc, char *argv[], void *userdata) { + pager_open(arg_no_pager, false); printf("%s [OPTIONS...] {COMMAND} ...\n\n" "Send control commands to or query the virtual machine and container\n" diff --git a/src/network/networkd-address.c b/src/network/networkd-address.c index 5498e352d8..ed52d5e42d 100644 --- a/src/network/networkd-address.c +++ b/src/network/networkd-address.c @@ -571,6 +571,21 @@ int address_configure( address->flags |= IFA_F_PERMANENT; + if (address->home_address) + address->flags |= IFA_F_HOMEADDRESS; + + if (address->duplicate_address_detection) + address->flags |= IFA_F_NODAD; + + if (address->manage_temporary_address) + address->flags |= IFA_F_MANAGETEMPADDR; + + if (address->prefix_route) + address->flags |= IFA_F_NOPREFIXROUTE; + + if (address->autojoin) + address->flags |= IFA_F_MCAUTOJOIN; + r = sd_rtnl_message_addr_set_flags(req, (address->flags & 0xff)); if (r < 0) return log_error_errno(r, "Could not set flags: %m"); @@ -856,6 +871,50 @@ int config_parse_lifetime(const char *unit, return 0; } +int config_parse_address_flags(const char *unit, + const char *filename, + unsigned line, + const char *section, + unsigned section_line, + const char *lvalue, + int ltype, + const char *rvalue, + void *data, + void *userdata) { + Network *network = userdata; + _cleanup_address_free_ Address *n = NULL; + int r; + + assert(filename); + assert(section); + assert(lvalue); + assert(rvalue); + assert(data); + + r = address_new_static(network, section_line, &n); + if (r < 0) + return r; + + r = parse_boolean(rvalue); + if (r < 0) { + log_syntax(unit, LOG_ERR, filename, line, r, "Failed to parse address flag, ignoring: %s", rvalue); + return 0; + } + + if (streq(lvalue, "HomeAddress")) + n->home_address = r; + else if (streq(lvalue, "DuplicateAddressDetection")) + n->duplicate_address_detection = r; + else if (streq(lvalue, "ManageTemporaryAddress")) + n->manage_temporary_address = r; + else if (streq(lvalue, "PrefixRoute")) + n->prefix_route = r; + else if (streq(lvalue, "AutoJoin")) + n->autojoin = r; + + return 0; +} + bool address_is_ready(const Address *a) { assert(a); diff --git a/src/network/networkd-address.h b/src/network/networkd-address.h index 03c4bea7c6..bc3b4fc7f3 100644 --- a/src/network/networkd-address.h +++ b/src/network/networkd-address.h @@ -53,6 +53,11 @@ struct Address { union in_addr_union in_addr_peer; bool ip_masquerade_done:1; + bool duplicate_address_detection; + bool manage_temporary_address; + bool home_address; + bool prefix_route; + bool autojoin; LIST_FIELDS(Address, addresses); }; @@ -77,3 +82,4 @@ int config_parse_address(const char *unit, const char *filename, unsigned line, int config_parse_broadcast(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata); int config_parse_label(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata); int config_parse_lifetime(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata); +int config_parse_address_flags(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata); diff --git a/src/network/networkd-netdev-gperf.gperf b/src/network/networkd-netdev-gperf.gperf index 6b7ab7ab87..323eaa8032 100644 --- a/src/network/networkd-netdev-gperf.gperf +++ b/src/network/networkd-netdev-gperf.gperf @@ -63,10 +63,13 @@ VXLAN.L2MissNotification, config_parse_bool, 0, VXLAN.L3MissNotification, config_parse_bool, 0, offsetof(VxLan, l3miss) VXLAN.RouteShortCircuit, config_parse_bool, 0, offsetof(VxLan, route_short_circuit) VXLAN.UDPCheckSum, config_parse_bool, 0, offsetof(VxLan, udpcsum) +VXLAN.UDPChecksum, config_parse_bool, 0, offsetof(VxLan, udpcsum) VXLAN.UDP6ZeroCheckSumRx, config_parse_bool, 0, offsetof(VxLan, udp6zerocsumrx) +VXLAN.UDP6ZeroChecksumRx, config_parse_bool, 0, offsetof(VxLan, udp6zerocsumrx) VXLAN.UDP6ZeroCheckSumTx, config_parse_bool, 0, offsetof(VxLan, udp6zerocsumtx) -VXLAN.RemoteCheckSumTx, config_parse_bool, 0, offsetof(VxLan, remote_csum_tx) -VXLAN.RemoteCheckSumRx, config_parse_bool, 0, offsetof(VxLan, remote_csum_rx) +VXLAN.UDP6ZeroChecksumTx, config_parse_bool, 0, offsetof(VxLan, udp6zerocsumtx) +VXLAN.RemoteChecksumTx, config_parse_bool, 0, offsetof(VxLan, remote_csum_tx) +VXLAN.RemoteChecksumRx, config_parse_bool, 0, offsetof(VxLan, remote_csum_rx) VXLAN.FDBAgeingSec, config_parse_sec, 0, offsetof(VxLan, fdb_ageing) VXLAN.GroupPolicyExtension, config_parse_bool, 0, offsetof(VxLan, group_policy) VXLAN.MaximumFDBEntries, config_parse_unsigned, 0, offsetof(VxLan, max_fdb) diff --git a/src/network/networkd-network-gperf.gperf b/src/network/networkd-network-gperf.gperf index 62779c7c48..5587961b9f 100644 --- a/src/network/networkd-network-gperf.gperf +++ b/src/network/networkd-network-gperf.gperf @@ -70,6 +70,11 @@ Address.Peer, config_parse_address, Address.Broadcast, config_parse_broadcast, 0, 0 Address.Label, config_parse_label, 0, 0 Address.PreferredLifetime, config_parse_lifetime, 0, 0 +Address.HomeAddress, config_parse_address_flags, 0, 0 +Address.DuplicateAddressDetection, config_parse_address_flags, 0, 0 +Address.ManageTemporaryAddress, config_parse_address_flags, 0, 0 +Address.PrefixRoute, config_parse_address_flags, 0, 0 +Address.AutoJoin, config_parse_address_flags, 0, 0 Route.Gateway, config_parse_gateway, 0, 0 Route.Destination, config_parse_destination, 0, 0 Route.Source, config_parse_destination, 0, 0 diff --git a/src/nspawn/nspawn.c b/src/nspawn/nspawn.c index a08377b3a3..a173d171e1 100644 --- a/src/nspawn/nspawn.c +++ b/src/nspawn/nspawn.c @@ -316,17 +316,10 @@ static int custom_mounts_prepare(void) { return 0; } -static int detect_unified_cgroup_hierarchy(void) { +static int detect_unified_cgroup_hierarchy(const char *directory) { const char *e; int r, all_unified, systemd_unified; - all_unified = cg_all_unified(); - systemd_unified = cg_unified(SYSTEMD_CGROUP_CONTROLLER); - - if (all_unified < 0 || systemd_unified < 0) - return log_error_errno(all_unified < 0 ? all_unified : systemd_unified, - "Failed to determine whether the unified cgroups hierarchy is used: %m"); - /* Allow the user to control whether the unified hierarchy is used */ e = getenv("UNIFIED_CGROUP_HIERARCHY"); if (e) { @@ -341,12 +334,34 @@ static int detect_unified_cgroup_hierarchy(void) { return 0; } + all_unified = cg_all_unified(); + systemd_unified = cg_unified(SYSTEMD_CGROUP_CONTROLLER); + + if (all_unified < 0 || systemd_unified < 0) + return log_error_errno(all_unified < 0 ? all_unified : systemd_unified, + "Failed to determine whether the unified cgroups hierarchy is used: %m"); + /* Otherwise inherit the default from the host system */ - if (all_unified > 0) - arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_ALL; - else if (systemd_unified > 0) - arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_SYSTEMD; - else + if (all_unified > 0) { + /* Unified cgroup hierarchy support was added in 230. Unfortunately the detection + * routine only detects 231, so we'll have a false negative here for 230. */ + r = systemd_installation_has_version(directory, 230); + if (r < 0) + return log_error_errno(r, "Failed to determine systemd version in container: %m"); + if (r > 0) + arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_ALL; + else + arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_NONE; + } else if (systemd_unified > 0) { + /* Mixed cgroup hierarchy support was added in 232 */ + r = systemd_installation_has_version(directory, 232); + if (r < 0) + return log_error_errno(r, "Failed to determine systemd version in container: %m"); + if (r > 0) + arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_SYSTEMD; + else + arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_NONE; + } else arg_unified_cgroup_hierarchy = CGROUP_UNIFIED_NONE; return 0; @@ -400,52 +415,52 @@ static int parse_argv(int argc, char *argv[]) { }; static const struct option options[] = { - { "help", no_argument, NULL, 'h' }, - { "version", no_argument, NULL, ARG_VERSION }, - { "directory", required_argument, NULL, 'D' }, - { "template", required_argument, NULL, ARG_TEMPLATE }, - { "ephemeral", no_argument, NULL, 'x' }, - { "user", required_argument, NULL, 'u' }, - { "private-network", no_argument, NULL, ARG_PRIVATE_NETWORK }, - { "as-pid2", no_argument, NULL, 'a' }, - { "boot", no_argument, NULL, 'b' }, - { "uuid", required_argument, NULL, ARG_UUID }, - { "read-only", no_argument, NULL, ARG_READ_ONLY }, - { "capability", required_argument, NULL, ARG_CAPABILITY }, - { "drop-capability", required_argument, NULL, ARG_DROP_CAPABILITY }, - { "link-journal", required_argument, NULL, ARG_LINK_JOURNAL }, - { "bind", required_argument, NULL, ARG_BIND }, - { "bind-ro", required_argument, NULL, ARG_BIND_RO }, - { "tmpfs", required_argument, NULL, ARG_TMPFS }, - { "overlay", required_argument, NULL, ARG_OVERLAY }, - { "overlay-ro", required_argument, NULL, ARG_OVERLAY_RO }, - { "machine", required_argument, NULL, 'M' }, - { "slice", required_argument, NULL, 'S' }, - { "setenv", required_argument, NULL, 'E' }, - { "selinux-context", required_argument, NULL, 'Z' }, - { "selinux-apifs-context", required_argument, NULL, 'L' }, - { "quiet", no_argument, NULL, 'q' }, - { "share-system", no_argument, NULL, ARG_SHARE_SYSTEM }, /* not documented */ - { "register", required_argument, NULL, ARG_REGISTER }, - { "keep-unit", no_argument, NULL, ARG_KEEP_UNIT }, - { "network-interface", required_argument, NULL, ARG_NETWORK_INTERFACE }, - { "network-macvlan", required_argument, NULL, ARG_NETWORK_MACVLAN }, - { "network-ipvlan", required_argument, NULL, ARG_NETWORK_IPVLAN }, - { "network-veth", no_argument, NULL, 'n' }, - { "network-veth-extra", required_argument, NULL, ARG_NETWORK_VETH_EXTRA}, - { "network-bridge", required_argument, NULL, ARG_NETWORK_BRIDGE }, - { "network-zone", required_argument, NULL, ARG_NETWORK_ZONE }, - { "personality", required_argument, NULL, ARG_PERSONALITY }, - { "image", required_argument, NULL, 'i' }, - { "volatile", optional_argument, NULL, ARG_VOLATILE }, - { "port", required_argument, NULL, 'p' }, - { "property", required_argument, NULL, ARG_PROPERTY }, - { "private-users", optional_argument, NULL, ARG_PRIVATE_USERS }, - { "private-users-chown", optional_argument, NULL, ARG_PRIVATE_USERS_CHOWN}, - { "kill-signal", required_argument, NULL, ARG_KILL_SIGNAL }, - { "settings", required_argument, NULL, ARG_SETTINGS }, - { "chdir", required_argument, NULL, ARG_CHDIR }, - { "notify-ready", required_argument, NULL, ARG_NOTIFY_READY }, + { "help", no_argument, NULL, 'h' }, + { "version", no_argument, NULL, ARG_VERSION }, + { "directory", required_argument, NULL, 'D' }, + { "template", required_argument, NULL, ARG_TEMPLATE }, + { "ephemeral", no_argument, NULL, 'x' }, + { "user", required_argument, NULL, 'u' }, + { "private-network", no_argument, NULL, ARG_PRIVATE_NETWORK }, + { "as-pid2", no_argument, NULL, 'a' }, + { "boot", no_argument, NULL, 'b' }, + { "uuid", required_argument, NULL, ARG_UUID }, + { "read-only", no_argument, NULL, ARG_READ_ONLY }, + { "capability", required_argument, NULL, ARG_CAPABILITY }, + { "drop-capability", required_argument, NULL, ARG_DROP_CAPABILITY }, + { "link-journal", required_argument, NULL, ARG_LINK_JOURNAL }, + { "bind", required_argument, NULL, ARG_BIND }, + { "bind-ro", required_argument, NULL, ARG_BIND_RO }, + { "tmpfs", required_argument, NULL, ARG_TMPFS }, + { "overlay", required_argument, NULL, ARG_OVERLAY }, + { "overlay-ro", required_argument, NULL, ARG_OVERLAY_RO }, + { "machine", required_argument, NULL, 'M' }, + { "slice", required_argument, NULL, 'S' }, + { "setenv", required_argument, NULL, 'E' }, + { "selinux-context", required_argument, NULL, 'Z' }, + { "selinux-apifs-context", required_argument, NULL, 'L' }, + { "quiet", no_argument, NULL, 'q' }, + { "share-system", no_argument, NULL, ARG_SHARE_SYSTEM }, /* not documented */ + { "register", required_argument, NULL, ARG_REGISTER }, + { "keep-unit", no_argument, NULL, ARG_KEEP_UNIT }, + { "network-interface", required_argument, NULL, ARG_NETWORK_INTERFACE }, + { "network-macvlan", required_argument, NULL, ARG_NETWORK_MACVLAN }, + { "network-ipvlan", required_argument, NULL, ARG_NETWORK_IPVLAN }, + { "network-veth", no_argument, NULL, 'n' }, + { "network-veth-extra", required_argument, NULL, ARG_NETWORK_VETH_EXTRA }, + { "network-bridge", required_argument, NULL, ARG_NETWORK_BRIDGE }, + { "network-zone", required_argument, NULL, ARG_NETWORK_ZONE }, + { "personality", required_argument, NULL, ARG_PERSONALITY }, + { "image", required_argument, NULL, 'i' }, + { "volatile", optional_argument, NULL, ARG_VOLATILE }, + { "port", required_argument, NULL, 'p' }, + { "property", required_argument, NULL, ARG_PROPERTY }, + { "private-users", optional_argument, NULL, ARG_PRIVATE_USERS }, + { "private-users-chown", optional_argument, NULL, ARG_PRIVATE_USERS_CHOWN }, + { "kill-signal", required_argument, NULL, ARG_KILL_SIGNAL }, + { "settings", required_argument, NULL, ARG_SETTINGS }, + { "chdir", required_argument, NULL, ARG_CHDIR }, + { "notify-ready", required_argument, NULL, ARG_NOTIFY_READY }, {} }; @@ -898,15 +913,21 @@ static int parse_argv(int argc, char *argv[]) { break; - case ARG_PRIVATE_USERS: + case ARG_PRIVATE_USERS: { + int boolean = -1; - r = optarg ? parse_boolean(optarg) : 1; - if (r == 0) { + if (!optarg) + boolean = true; + else if (!in_charset(optarg, DIGITS)) + /* do *not* parse numbers as booleans */ + boolean = parse_boolean(optarg); + + if (boolean == false) { /* no: User namespacing off */ arg_userns_mode = USER_NAMESPACE_NO; arg_uid_shift = UID_INVALID; arg_uid_range = UINT32_C(0x10000); - } else if (r > 0) { + } else if (boolean == true) { /* yes: User namespacing on, UID range is read from root dir */ arg_userns_mode = USER_NAMESPACE_FIXED; arg_uid_shift = UID_INVALID; @@ -930,23 +951,27 @@ static int parse_argv(int argc, char *argv[]) { shift = buffer; range++; - if (safe_atou32(range, &arg_uid_range) < 0 || arg_uid_range <= 0) { - log_error("Failed to parse UID range: %s", range); - return -EINVAL; - } + r = safe_atou32(range, &arg_uid_range); + if (r < 0) + return log_error_errno(r, "Failed to parse UID range \"%s\": %m", range); } else shift = optarg; - if (parse_uid(shift, &arg_uid_shift) < 0) { - log_error("Failed to parse UID: %s", optarg); - return -EINVAL; - } + r = parse_uid(shift, &arg_uid_shift); + if (r < 0) + return log_error_errno(r, "Failed to parse UID \"%s\": %m", optarg); arg_userns_mode = USER_NAMESPACE_FIXED; } + if (arg_uid_range <= 0) { + log_error("UID range cannot be 0."); + return -EINVAL; + } + arg_settings_mask |= SETTING_USERNS; break; + } case 'U': if (userns_supported()) { @@ -1125,10 +1150,6 @@ static int parse_argv(int argc, char *argv[]) { arg_caps_retain = (arg_caps_retain | plus | (arg_private_network ? 1ULL << CAP_NET_ADMIN : 0)) & ~minus; - r = detect_unified_cgroup_hierarchy(); - if (r < 0) - return r; - e = getenv("SYSTEMD_NSPAWN_CONTAINER_SERVICE"); if (e) arg_container_service_name = e; @@ -2970,6 +2991,10 @@ static int outer_child( if (r < 0) return r; + r = detect_unified_cgroup_hierarchy(directory); + if (r < 0) + return r; + if (arg_userns_mode != USER_NAMESPACE_NO) { /* Let the parent know which UID shift we read from the image */ l = send(uid_shift_socket, &arg_uid_shift, sizeof(arg_uid_shift), MSG_NOSIGNAL); @@ -3575,18 +3600,437 @@ static int load_settings(void) { return 0; } +static int run(int master, + const char* console, + const char *root_device, bool root_device_rw, + const char *home_device, bool home_device_rw, + const char *srv_device, bool srv_device_rw, + const char *esp_device, + bool interactive, + bool secondary, + FDSet *fds, + char veth_name[IFNAMSIZ], bool *veth_created, + union in_addr_union *exposed, + pid_t *pid, int *ret) { + + static const struct sigaction sa = { + .sa_handler = nop_signal_handler, + .sa_flags = SA_NOCLDSTOP, + }; + + _cleanup_release_lock_file_ LockFile uid_shift_lock = LOCK_FILE_INIT; + _cleanup_close_ int etc_passwd_lock = -1; + _cleanup_close_pair_ int + kmsg_socket_pair[2] = { -1, -1 }, + rtnl_socket_pair[2] = { -1, -1 }, + pid_socket_pair[2] = { -1, -1 }, + uuid_socket_pair[2] = { -1, -1 }, + notify_socket_pair[2] = { -1, -1 }, + uid_shift_socket_pair[2] = { -1, -1 }; + _cleanup_close_ int notify_socket= -1; + _cleanup_(barrier_destroy) Barrier barrier = BARRIER_NULL; + _cleanup_(sd_event_unrefp) sd_event *event = NULL; + _cleanup_(pty_forward_freep) PTYForward *forward = NULL; + _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL; + ContainerStatus container_status = 0; + char last_char = 0; + int ifi = 0, r; + ssize_t l; + sigset_t mask_chld; + + assert_se(sigemptyset(&mask_chld) == 0); + assert_se(sigaddset(&mask_chld, SIGCHLD) == 0); + + if (arg_userns_mode == USER_NAMESPACE_PICK) { + /* When we shall pick the UID/GID range, let's first lock /etc/passwd, so that we can safely + * check with getpwuid() if the specific user already exists. Note that /etc might be + * read-only, in which case this will fail with EROFS. But that's really OK, as in that case we + * can be reasonably sure that no users are going to be added. Note that getpwuid() checks are + * really just an extra safety net. We kinda assume that the UID range we allocate from is + * really ours. */ + + etc_passwd_lock = take_etc_passwd_lock(NULL); + if (etc_passwd_lock < 0 && etc_passwd_lock != -EROFS) + return log_error_errno(etc_passwd_lock, "Failed to take /etc/passwd lock: %m"); + } + + r = barrier_create(&barrier); + if (r < 0) + return log_error_errno(r, "Cannot initialize IPC barrier: %m"); + + if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, kmsg_socket_pair) < 0) + return log_error_errno(errno, "Failed to create kmsg socket pair: %m"); + + if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, rtnl_socket_pair) < 0) + return log_error_errno(errno, "Failed to create rtnl socket pair: %m"); + + if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, pid_socket_pair) < 0) + return log_error_errno(errno, "Failed to create pid socket pair: %m"); + + if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, uuid_socket_pair) < 0) + return log_error_errno(errno, "Failed to create id socket pair: %m"); + + if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, notify_socket_pair) < 0) + return log_error_errno(errno, "Failed to create notify socket pair: %m"); + + if (arg_userns_mode != USER_NAMESPACE_NO) + if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, uid_shift_socket_pair) < 0) + return log_error_errno(errno, "Failed to create uid shift socket pair: %m"); + + /* Child can be killed before execv(), so handle SIGCHLD in order to interrupt + * parent's blocking calls and give it a chance to call wait() and terminate. */ + r = sigprocmask(SIG_UNBLOCK, &mask_chld, NULL); + if (r < 0) + return log_error_errno(errno, "Failed to change the signal mask: %m"); + + r = sigaction(SIGCHLD, &sa, NULL); + if (r < 0) + return log_error_errno(errno, "Failed to install SIGCHLD handler: %m"); + + *pid = raw_clone(SIGCHLD|CLONE_NEWNS); + if (*pid < 0) + return log_error_errno(errno, "clone() failed%s: %m", + errno == EINVAL ? + ", do you have namespace support enabled in your kernel? (You need UTS, IPC, PID and NET namespacing built in)" : ""); + + if (*pid == 0) { + /* The outer child only has a file system namespace. */ + barrier_set_role(&barrier, BARRIER_CHILD); + + master = safe_close(master); + + kmsg_socket_pair[0] = safe_close(kmsg_socket_pair[0]); + rtnl_socket_pair[0] = safe_close(rtnl_socket_pair[0]); + pid_socket_pair[0] = safe_close(pid_socket_pair[0]); + uuid_socket_pair[0] = safe_close(uuid_socket_pair[0]); + notify_socket_pair[0] = safe_close(notify_socket_pair[0]); + uid_shift_socket_pair[0] = safe_close(uid_shift_socket_pair[0]); + + (void) reset_all_signal_handlers(); + (void) reset_signal_mask(); + + r = outer_child(&barrier, + arg_directory, + console, + root_device, root_device_rw, + home_device, home_device_rw, + srv_device, srv_device_rw, + esp_device, + interactive, + secondary, + pid_socket_pair[1], + uuid_socket_pair[1], + notify_socket_pair[1], + kmsg_socket_pair[1], + rtnl_socket_pair[1], + uid_shift_socket_pair[1], + fds); + if (r < 0) + _exit(EXIT_FAILURE); + + _exit(EXIT_SUCCESS); + } + + barrier_set_role(&barrier, BARRIER_PARENT); + + fds = fdset_free(fds); + + kmsg_socket_pair[1] = safe_close(kmsg_socket_pair[1]); + rtnl_socket_pair[1] = safe_close(rtnl_socket_pair[1]); + pid_socket_pair[1] = safe_close(pid_socket_pair[1]); + uuid_socket_pair[1] = safe_close(uuid_socket_pair[1]); + notify_socket_pair[1] = safe_close(notify_socket_pair[1]); + uid_shift_socket_pair[1] = safe_close(uid_shift_socket_pair[1]); + + if (arg_userns_mode != USER_NAMESPACE_NO) { + /* The child just let us know the UID shift it might have read from the image. */ + l = recv(uid_shift_socket_pair[0], &arg_uid_shift, sizeof arg_uid_shift, 0); + if (l < 0) + return log_error_errno(errno, "Failed to read UID shift: %m"); + + if (l != sizeof arg_uid_shift) { + log_error("Short read while reading UID shift."); + return -EIO; + } + + if (arg_userns_mode == USER_NAMESPACE_PICK) { + /* If we are supposed to pick the UID shift, let's try to use the shift read from the + * image, but if that's already in use, pick a new one, and report back to the child, + * which one we now picked. */ + + r = uid_shift_pick(&arg_uid_shift, &uid_shift_lock); + if (r < 0) + return log_error_errno(r, "Failed to pick suitable UID/GID range: %m"); + + l = send(uid_shift_socket_pair[0], &arg_uid_shift, sizeof arg_uid_shift, MSG_NOSIGNAL); + if (l < 0) + return log_error_errno(errno, "Failed to send UID shift: %m"); + if (l != sizeof arg_uid_shift) { + log_error("Short write while writing UID shift."); + return -EIO; + } + } + } + + /* Wait for the outer child. */ + r = wait_for_terminate_and_warn("namespace helper", *pid, NULL); + if (r != 0) + return r < 0 ? r : -EIO; + + /* And now retrieve the PID of the inner child. */ + l = recv(pid_socket_pair[0], pid, sizeof *pid, 0); + if (l < 0) + return log_error_errno(errno, "Failed to read inner child PID: %m"); + if (l != sizeof *pid) { + log_error("Short read while reading inner child PID."); + return -EIO; + } + + /* We also retrieve container UUID in case it was generated by outer child */ + l = recv(uuid_socket_pair[0], &arg_uuid, sizeof arg_uuid, 0); + if (l < 0) + return log_error_errno(errno, "Failed to read container machine ID: %m"); + if (l != sizeof(arg_uuid)) { + log_error("Short read while reading container machined ID."); + return -EIO; + } + + /* We also retrieve the socket used for notifications generated by outer child */ + notify_socket = receive_one_fd(notify_socket_pair[0], 0); + if (notify_socket < 0) + return log_error_errno(notify_socket, + "Failed to receive notification socket from the outer child: %m"); + + log_debug("Init process invoked as PID "PID_FMT, *pid); + + if (arg_userns_mode != USER_NAMESPACE_NO) { + if (!barrier_place_and_sync(&barrier)) { /* #1 */ + log_error("Child died too early."); + return -ESRCH; + } + + r = setup_uid_map(*pid); + if (r < 0) + return r; + + (void) barrier_place(&barrier); /* #2 */ + } + + if (arg_private_network) { + + r = move_network_interfaces(*pid, arg_network_interfaces); + if (r < 0) + return r; + + if (arg_network_veth) { + r = setup_veth(arg_machine, *pid, veth_name, + arg_network_bridge || arg_network_zone); + if (r < 0) + return r; + else if (r > 0) + ifi = r; + + if (arg_network_bridge) { + /* Add the interface to a bridge */ + r = setup_bridge(veth_name, arg_network_bridge, false); + if (r < 0) + return r; + if (r > 0) + ifi = r; + } else if (arg_network_zone) { + /* Add the interface to a bridge, possibly creating it */ + r = setup_bridge(veth_name, arg_network_zone, true); + if (r < 0) + return r; + if (r > 0) + ifi = r; + } + } + + r = setup_veth_extra(arg_machine, *pid, arg_network_veth_extra); + if (r < 0) + return r; + + /* We created the primary and extra veth links now; let's remember this, so that we know to + remove them later on. Note that we don't bother with removing veth links that were created + here when their setup failed half-way, because in that case the kernel should be able to + remove them on its own, since they cannot be referenced by anything yet. */ + *veth_created = true; + + r = setup_macvlan(arg_machine, *pid, arg_network_macvlan); + if (r < 0) + return r; + + r = setup_ipvlan(arg_machine, *pid, arg_network_ipvlan); + if (r < 0) + return r; + } + + if (arg_register) { + r = register_machine( + arg_machine, + *pid, + arg_directory, + arg_uuid, + ifi, + arg_slice, + arg_custom_mounts, arg_n_custom_mounts, + arg_kill_signal, + arg_property, + arg_keep_unit, + arg_container_service_name); + if (r < 0) + return r; + } + + r = sync_cgroup(*pid, arg_unified_cgroup_hierarchy); + if (r < 0) + return r; + + if (arg_keep_unit) { + r = create_subcgroup(*pid, arg_unified_cgroup_hierarchy); + if (r < 0) + return r; + } + + r = chown_cgroup(*pid, arg_uid_shift); + if (r < 0) + return r; + + /* Notify the child that the parent is ready with all + * its setup (including cgroup-ification), and that + * the child can now hand over control to the code to + * run inside the container. */ + (void) barrier_place(&barrier); /* #3 */ + + /* Block SIGCHLD here, before notifying child. + * process_pty() will handle it with the other signals. */ + assert_se(sigprocmask(SIG_BLOCK, &mask_chld, NULL) >= 0); + + /* Reset signal to default */ + r = default_signals(SIGCHLD, -1); + if (r < 0) + return log_error_errno(r, "Failed to reset SIGCHLD: %m"); + + r = sd_event_new(&event); + if (r < 0) + return log_error_errno(r, "Failed to get default event source: %m"); + + r = setup_sd_notify_parent(event, notify_socket, PID_TO_PTR(*pid)); + if (r < 0) + return r; + + /* Let the child know that we are ready and wait that the child is completely ready now. */ + if (!barrier_place_and_sync(&barrier)) { /* #4 */ + log_error("Child died too early."); + return -ESRCH; + } + + /* At this point we have made use of the UID we picked, and thus nss-mymachines + * will make them appear in getpwuid(), thus we can release the /etc/passwd lock. */ + etc_passwd_lock = safe_close(etc_passwd_lock); + + sd_notifyf(false, + "STATUS=Container running.\n" + "X_NSPAWN_LEADER_PID=" PID_FMT, *pid); + if (!arg_notify_ready) + sd_notify(false, "READY=1\n"); + + if (arg_kill_signal > 0) { + /* Try to kill the init system on SIGINT or SIGTERM */ + sd_event_add_signal(event, NULL, SIGINT, on_orderly_shutdown, PID_TO_PTR(*pid)); + sd_event_add_signal(event, NULL, SIGTERM, on_orderly_shutdown, PID_TO_PTR(*pid)); + } else { + /* Immediately exit */ + sd_event_add_signal(event, NULL, SIGINT, NULL, NULL); + sd_event_add_signal(event, NULL, SIGTERM, NULL, NULL); + } + + /* simply exit on sigchld */ + sd_event_add_signal(event, NULL, SIGCHLD, NULL, NULL); + + if (arg_expose_ports) { + r = expose_port_watch_rtnl(event, rtnl_socket_pair[0], on_address_change, exposed, &rtnl); + if (r < 0) + return r; + + (void) expose_port_execute(rtnl, arg_expose_ports, exposed); + } + + rtnl_socket_pair[0] = safe_close(rtnl_socket_pair[0]); + + r = pty_forward_new(event, master, + PTY_FORWARD_IGNORE_VHANGUP | (interactive ? 0 : PTY_FORWARD_READ_ONLY), + &forward); + if (r < 0) + return log_error_errno(r, "Failed to create PTY forwarder: %m"); + + r = sd_event_loop(event); + if (r < 0) + return log_error_errno(r, "Failed to run event loop: %m"); + + pty_forward_get_last_char(forward, &last_char); + + forward = pty_forward_free(forward); + + if (!arg_quiet && last_char != '\n') + putc('\n', stdout); + + /* Kill if it is not dead yet anyway */ + if (arg_register && !arg_keep_unit) + terminate_machine(*pid); + + /* Normally redundant, but better safe than sorry */ + kill(*pid, SIGKILL); + + r = wait_for_container(*pid, &container_status); + *pid = 0; + + if (r < 0) + /* We failed to wait for the container, or the container exited abnormally. */ + return r; + if (r > 0 || container_status == CONTAINER_TERMINATED) { + /* r > 0 → The container exited with a non-zero status. + * As a special case, we need to replace 133 with a different value, + * because 133 is special-cased in the service file to reboot the container. + * otherwise → The container exited with zero status and a reboot was not requested. + */ + if (r == 133) + r = EXIT_FAILURE; /* replace 133 with the general failure code */ + *ret = r; + return 0; /* finito */ + } + + /* CONTAINER_REBOOTED, loop again */ + + if (arg_keep_unit) { + /* Special handling if we are running as a service: instead of simply + * restarting the machine we want to restart the entire service, so let's + * inform systemd about this with the special exit code 133. The service + * file uses RestartForceExitStatus=133 so that this results in a full + * nspawn restart. This is necessary since we might have cgroup parameters + * set we want to have flushed out. */ + *ret = 0; + return 133; + } + + expose_port_flush(arg_expose_ports, exposed); + + (void) remove_veth_links(veth_name, arg_network_veth_extra); + *veth_created = false; + return 1; /* loop again */ +} + int main(int argc, char *argv[]) { _cleanup_free_ char *device_path = NULL, *root_device = NULL, *home_device = NULL, *srv_device = NULL, *esp_device = NULL, *console = NULL; bool root_device_rw = true, home_device_rw = true, srv_device_rw = true; _cleanup_close_ int master = -1, image_fd = -1; _cleanup_fdset_free_ FDSet *fds = NULL; - int r, n_fd_passed, loop_nr = -1; + int r, n_fd_passed, loop_nr = -1, ret = EXIT_FAILURE; char veth_name[IFNAMSIZ] = ""; bool secondary = false, remove_subvol = false; - sigset_t mask_chld; pid_t pid = 0; - int ret = EXIT_SUCCESS; union in_addr_union exposed = {}; _cleanup_release_lock_file_ LockFile tree_global_lock = LOCK_FILE_INIT, tree_local_lock = LOCK_FILE_INIT; bool interactive, veth_created = false; @@ -3802,470 +4246,25 @@ int main(int argc, char *argv[]) { assert_se(sigprocmask_many(SIG_BLOCK, NULL, SIGCHLD, SIGWINCH, SIGTERM, SIGINT, -1) >= 0); - assert_se(sigemptyset(&mask_chld) == 0); - assert_se(sigaddset(&mask_chld, SIGCHLD) == 0); - if (prctl(PR_SET_CHILD_SUBREAPER, 1) < 0) { r = log_error_errno(errno, "Failed to become subreaper: %m"); goto finish; } for (;;) { - static const struct sigaction sa = { - .sa_handler = nop_signal_handler, - .sa_flags = SA_NOCLDSTOP, - }; - - _cleanup_release_lock_file_ LockFile uid_shift_lock = LOCK_FILE_INIT; - _cleanup_close_ int etc_passwd_lock = -1; - _cleanup_close_pair_ int - kmsg_socket_pair[2] = { -1, -1 }, - rtnl_socket_pair[2] = { -1, -1 }, - pid_socket_pair[2] = { -1, -1 }, - uuid_socket_pair[2] = { -1, -1 }, - notify_socket_pair[2] = { -1, -1 }, - uid_shift_socket_pair[2] = { -1, -1 }; - _cleanup_close_ int notify_socket= -1; - _cleanup_(barrier_destroy) Barrier barrier = BARRIER_NULL; - _cleanup_(sd_event_unrefp) sd_event *event = NULL; - _cleanup_(pty_forward_freep) PTYForward *forward = NULL; - _cleanup_(sd_netlink_unrefp) sd_netlink *rtnl = NULL; - ContainerStatus container_status = 0; - char last_char = 0; - int ifi = 0; - ssize_t l; - - if (arg_userns_mode == USER_NAMESPACE_PICK) { - /* When we shall pick the UID/GID range, let's first lock /etc/passwd, so that we can safely - * check with getpwuid() if the specific user already exists. Note that /etc might be - * read-only, in which case this will fail with EROFS. But that's really OK, as in that case we - * can be reasonably sure that no users are going to be added. Note that getpwuid() checks are - * really just an extra safety net. We kinda assume that the UID range we allocate from is - * really ours. */ - - etc_passwd_lock = take_etc_passwd_lock(NULL); - if (etc_passwd_lock < 0 && etc_passwd_lock != -EROFS) { - log_error_errno(r, "Failed to take /etc/passwd lock: %m"); - goto finish; - } - } - - r = barrier_create(&barrier); - if (r < 0) { - log_error_errno(r, "Cannot initialize IPC barrier: %m"); - goto finish; - } - - if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, kmsg_socket_pair) < 0) { - r = log_error_errno(errno, "Failed to create kmsg socket pair: %m"); - goto finish; - } - - if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, rtnl_socket_pair) < 0) { - r = log_error_errno(errno, "Failed to create rtnl socket pair: %m"); - goto finish; - } - - if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, pid_socket_pair) < 0) { - r = log_error_errno(errno, "Failed to create pid socket pair: %m"); - goto finish; - } - - if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, uuid_socket_pair) < 0) { - r = log_error_errno(errno, "Failed to create id socket pair: %m"); - goto finish; - } - - if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, notify_socket_pair) < 0) { - r = log_error_errno(errno, "Failed to create notify socket pair: %m"); - goto finish; - } - - if (arg_userns_mode != USER_NAMESPACE_NO) - if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, uid_shift_socket_pair) < 0) { - r = log_error_errno(errno, "Failed to create uid shift socket pair: %m"); - goto finish; - } - - /* Child can be killed before execv(), so handle SIGCHLD - * in order to interrupt parent's blocking calls and - * give it a chance to call wait() and terminate. */ - r = sigprocmask(SIG_UNBLOCK, &mask_chld, NULL); - if (r < 0) { - r = log_error_errno(errno, "Failed to change the signal mask: %m"); - goto finish; - } - - r = sigaction(SIGCHLD, &sa, NULL); - if (r < 0) { - r = log_error_errno(errno, "Failed to install SIGCHLD handler: %m"); - goto finish; - } - - pid = raw_clone(SIGCHLD|CLONE_NEWNS); - if (pid < 0) { - if (errno == EINVAL) - r = log_error_errno(errno, "clone() failed, do you have namespace support enabled in your kernel? (You need UTS, IPC, PID and NET namespacing built in): %m"); - else - r = log_error_errno(errno, "clone() failed: %m"); - - goto finish; - } - - if (pid == 0) { - /* The outer child only has a file system namespace. */ - barrier_set_role(&barrier, BARRIER_CHILD); - - master = safe_close(master); - - kmsg_socket_pair[0] = safe_close(kmsg_socket_pair[0]); - rtnl_socket_pair[0] = safe_close(rtnl_socket_pair[0]); - pid_socket_pair[0] = safe_close(pid_socket_pair[0]); - uuid_socket_pair[0] = safe_close(uuid_socket_pair[0]); - notify_socket_pair[0] = safe_close(notify_socket_pair[0]); - uid_shift_socket_pair[0] = safe_close(uid_shift_socket_pair[0]); - - (void) reset_all_signal_handlers(); - (void) reset_signal_mask(); - - r = outer_child(&barrier, - arg_directory, - console, - root_device, root_device_rw, - home_device, home_device_rw, - srv_device, srv_device_rw, - esp_device, - interactive, - secondary, - pid_socket_pair[1], - uuid_socket_pair[1], - notify_socket_pair[1], - kmsg_socket_pair[1], - rtnl_socket_pair[1], - uid_shift_socket_pair[1], - fds); - if (r < 0) - _exit(EXIT_FAILURE); - - _exit(EXIT_SUCCESS); - } - - barrier_set_role(&barrier, BARRIER_PARENT); - - fds = fdset_free(fds); - - kmsg_socket_pair[1] = safe_close(kmsg_socket_pair[1]); - rtnl_socket_pair[1] = safe_close(rtnl_socket_pair[1]); - pid_socket_pair[1] = safe_close(pid_socket_pair[1]); - uuid_socket_pair[1] = safe_close(uuid_socket_pair[1]); - notify_socket_pair[1] = safe_close(notify_socket_pair[1]); - uid_shift_socket_pair[1] = safe_close(uid_shift_socket_pair[1]); - - if (arg_userns_mode != USER_NAMESPACE_NO) { - /* The child just let us know the UID shift it might have read from the image. */ - l = recv(uid_shift_socket_pair[0], &arg_uid_shift, sizeof(arg_uid_shift), 0); - if (l < 0) { - r = log_error_errno(errno, "Failed to read UID shift: %m"); - goto finish; - } - if (l != sizeof(arg_uid_shift)) { - log_error("Short read while reading UID shift."); - r = EIO; - goto finish; - } - - if (arg_userns_mode == USER_NAMESPACE_PICK) { - /* If we are supposed to pick the UID shift, let's try to use the shift read from the - * image, but if that's already in use, pick a new one, and report back to the child, - * which one we now picked. */ - - r = uid_shift_pick(&arg_uid_shift, &uid_shift_lock); - if (r < 0) { - log_error_errno(r, "Failed to pick suitable UID/GID range: %m"); - goto finish; - } - - l = send(uid_shift_socket_pair[0], &arg_uid_shift, sizeof(arg_uid_shift), MSG_NOSIGNAL); - if (l < 0) { - r = log_error_errno(errno, "Failed to send UID shift: %m"); - goto finish; - } - if (l != sizeof(arg_uid_shift)) { - log_error("Short write while writing UID shift."); - r = -EIO; - goto finish; - } - } - } - - /* Wait for the outer child. */ - r = wait_for_terminate_and_warn("namespace helper", pid, NULL); - if (r < 0) - goto finish; - if (r != 0) { - r = -EIO; - goto finish; - } - pid = 0; - - /* And now retrieve the PID of the inner child. */ - l = recv(pid_socket_pair[0], &pid, sizeof(pid), 0); - if (l < 0) { - r = log_error_errno(errno, "Failed to read inner child PID: %m"); - goto finish; - } - if (l != sizeof(pid)) { - log_error("Short read while reading inner child PID."); - r = EIO; - goto finish; - } - - /* We also retrieve container UUID in case it was generated by outer child */ - l = recv(uuid_socket_pair[0], &arg_uuid, sizeof(arg_uuid), 0); - if (l < 0) { - r = log_error_errno(errno, "Failed to read container machine ID: %m"); - goto finish; - } - if (l != sizeof(arg_uuid)) { - log_error("Short read while reading container machined ID."); - r = EIO; - goto finish; - } - - /* We also retrieve the socket used for notifications generated by outer child */ - notify_socket = receive_one_fd(notify_socket_pair[0], 0); - if (notify_socket < 0) { - r = log_error_errno(errno, "Failed to receive notification socket from the outer child: %m"); - goto finish; - } - - log_debug("Init process invoked as PID " PID_FMT, pid); - - if (arg_userns_mode != USER_NAMESPACE_NO) { - if (!barrier_place_and_sync(&barrier)) { /* #1 */ - log_error("Child died too early."); - r = -ESRCH; - goto finish; - } - - r = setup_uid_map(pid); - if (r < 0) - goto finish; - - (void) barrier_place(&barrier); /* #2 */ - } - - if (arg_private_network) { - - r = move_network_interfaces(pid, arg_network_interfaces); - if (r < 0) - goto finish; - - if (arg_network_veth) { - r = setup_veth(arg_machine, pid, veth_name, - arg_network_bridge || arg_network_zone); - if (r < 0) - goto finish; - else if (r > 0) - ifi = r; - - if (arg_network_bridge) { - /* Add the interface to a bridge */ - r = setup_bridge(veth_name, arg_network_bridge, false); - if (r < 0) - goto finish; - if (r > 0) - ifi = r; - } else if (arg_network_zone) { - /* Add the interface to a bridge, possibly creating it */ - r = setup_bridge(veth_name, arg_network_zone, true); - if (r < 0) - goto finish; - if (r > 0) - ifi = r; - } - } - - r = setup_veth_extra(arg_machine, pid, arg_network_veth_extra); - if (r < 0) - goto finish; - - /* We created the primary and extra veth links now; let's remember this, so that we know to - remove them later on. Note that we don't bother with removing veth links that were created - here when their setup failed half-way, because in that case the kernel should be able to - remove them on its own, since they cannot be referenced by anything yet. */ - veth_created = true; - - r = setup_macvlan(arg_machine, pid, arg_network_macvlan); - if (r < 0) - goto finish; - - r = setup_ipvlan(arg_machine, pid, arg_network_ipvlan); - if (r < 0) - goto finish; - } - - if (arg_register) { - r = register_machine( - arg_machine, - pid, - arg_directory, - arg_uuid, - ifi, - arg_slice, - arg_custom_mounts, arg_n_custom_mounts, - arg_kill_signal, - arg_property, - arg_keep_unit, - arg_container_service_name); - if (r < 0) - goto finish; - } - - r = sync_cgroup(pid, arg_unified_cgroup_hierarchy); - if (r < 0) - goto finish; - - if (arg_keep_unit) { - r = create_subcgroup(pid, arg_unified_cgroup_hierarchy); - if (r < 0) - goto finish; - } - - r = chown_cgroup(pid, arg_uid_shift); - if (r < 0) - goto finish; - - /* Notify the child that the parent is ready with all - * its setup (including cgroup-ification), and that - * the child can now hand over control to the code to - * run inside the container. */ - (void) barrier_place(&barrier); /* #3 */ - - /* Block SIGCHLD here, before notifying child. - * process_pty() will handle it with the other signals. */ - assert_se(sigprocmask(SIG_BLOCK, &mask_chld, NULL) >= 0); - - /* Reset signal to default */ - r = default_signals(SIGCHLD, -1); - if (r < 0) { - log_error_errno(r, "Failed to reset SIGCHLD: %m"); - goto finish; - } - - r = sd_event_new(&event); - if (r < 0) { - log_error_errno(r, "Failed to get default event source: %m"); - goto finish; - } - - r = setup_sd_notify_parent(event, notify_socket, PID_TO_PTR(pid)); - if (r < 0) - goto finish; - - /* Let the child know that we are ready and wait that the child is completely ready now. */ - if (!barrier_place_and_sync(&barrier)) { /* #4 */ - log_error("Child died too early."); - r = -ESRCH; - goto finish; - } - - /* At this point we have made use of the UID we picked, and thus nss-mymachines will make them appear - * in getpwuid(), thus we can release the /etc/passwd lock. */ - etc_passwd_lock = safe_close(etc_passwd_lock); - - sd_notifyf(false, - "STATUS=Container running.\n" - "X_NSPAWN_LEADER_PID=" PID_FMT, pid); - if (!arg_notify_ready) - sd_notify(false, "READY=1\n"); - - if (arg_kill_signal > 0) { - /* Try to kill the init system on SIGINT or SIGTERM */ - sd_event_add_signal(event, NULL, SIGINT, on_orderly_shutdown, PID_TO_PTR(pid)); - sd_event_add_signal(event, NULL, SIGTERM, on_orderly_shutdown, PID_TO_PTR(pid)); - } else { - /* Immediately exit */ - sd_event_add_signal(event, NULL, SIGINT, NULL, NULL); - sd_event_add_signal(event, NULL, SIGTERM, NULL, NULL); - } - - /* simply exit on sigchld */ - sd_event_add_signal(event, NULL, SIGCHLD, NULL, NULL); - - if (arg_expose_ports) { - r = expose_port_watch_rtnl(event, rtnl_socket_pair[0], on_address_change, &exposed, &rtnl); - if (r < 0) - goto finish; - - (void) expose_port_execute(rtnl, arg_expose_ports, &exposed); - } - - rtnl_socket_pair[0] = safe_close(rtnl_socket_pair[0]); - - r = pty_forward_new(event, master, PTY_FORWARD_IGNORE_VHANGUP | (interactive ? 0 : PTY_FORWARD_READ_ONLY), &forward); - if (r < 0) { - log_error_errno(r, "Failed to create PTY forwarder: %m"); - goto finish; - } - - r = sd_event_loop(event); - if (r < 0) { - log_error_errno(r, "Failed to run event loop: %m"); - goto finish; - } - - pty_forward_get_last_char(forward, &last_char); - - forward = pty_forward_free(forward); - - if (!arg_quiet && last_char != '\n') - putc('\n', stdout); - - /* Kill if it is not dead yet anyway */ - if (arg_register && !arg_keep_unit) - terminate_machine(pid); - - /* Normally redundant, but better safe than sorry */ - kill(pid, SIGKILL); - - r = wait_for_container(pid, &container_status); - pid = 0; - - if (r < 0) - /* We failed to wait for the container, or the - * container exited abnormally */ - goto finish; - else if (r > 0 || container_status == CONTAINER_TERMINATED) { - /* The container exited with a non-zero - * status, or with zero status and no reboot - * was requested. */ - ret = r; - break; - } - - /* CONTAINER_REBOOTED, loop again */ - - if (arg_keep_unit) { - /* Special handling if we are running as a - * service: instead of simply restarting the - * machine we want to restart the entire - * service, so let's inform systemd about this - * with the special exit code 133. The service - * file uses RestartForceExitStatus=133 so - * that this results in a full nspawn - * restart. This is necessary since we might - * have cgroup parameters set we want to have - * flushed out. */ - ret = 133; - r = 0; + r = run(master, + console, + root_device, root_device_rw, + home_device, home_device_rw, + srv_device, srv_device_rw, + esp_device, + interactive, secondary, + fds, + veth_name, &veth_created, + &exposed, + &pid, &ret); + if (r <= 0) break; - } - - expose_port_flush(arg_expose_ports, &exposed); - - (void) remove_veth_links(veth_name, arg_network_veth_extra); - veth_created = false; } finish: diff --git a/src/remount-fs/remount-fs.c b/src/remount-fs/remount-fs.c index 6468d1eecd..c3bdcaf1da 100644 --- a/src/remount-fs/remount-fs.c +++ b/src/remount-fs/remount-fs.c @@ -137,7 +137,7 @@ int main(int argc, char *argv[]) { s = hashmap_remove(pids, PID_TO_PTR(si.si_pid)); if (s) { - if (!is_clean_exit(si.si_code, si.si_status, NULL)) { + if (!is_clean_exit(si.si_code, si.si_status, EXIT_CLEAN_COMMAND, NULL)) { if (si.si_code == CLD_EXITED) log_error(MOUNT_PATH " for %s exited with exit status %i.", s, si.si_status); else diff --git a/src/resolve/resolve-tool.c b/src/resolve/resolve-tool.c index 07e4cd7d1d..8aa79049b6 100644 --- a/src/resolve/resolve-tool.c +++ b/src/resolve/resolve-tool.c @@ -395,7 +395,7 @@ static int output_rr_packet(const void *d, size_t l, int ifindex) { return 0; } -static int resolve_record(sd_bus *bus, const char *name, uint16_t class, uint16_t type) { +static int resolve_record(sd_bus *bus, const char *name, uint16_t class, uint16_t type, bool warn_missing) { _cleanup_(sd_bus_message_unrefp) sd_bus_message *req = NULL, *reply = NULL; _cleanup_(sd_bus_error_free) sd_bus_error error = SD_BUS_ERROR_NULL; char ifname[IF_NAMESIZE] = ""; @@ -430,7 +430,8 @@ static int resolve_record(sd_bus *bus, const char *name, uint16_t class, uint16_ r = sd_bus_call(bus, req, DNS_CALL_TIMEOUT_USEC, &error, &reply); if (r < 0) { - log_error("%s: resolve call failed: %s", name, bus_error_message(&error, r)); + if (warn_missing || r != -ENXIO) + log_error("%s: resolve call failed: %s", name, bus_error_message(&error, r)); return r; } @@ -488,7 +489,8 @@ static int resolve_record(sd_bus *bus, const char *name, uint16_t class, uint16_ return bus_log_parse_error(r); if (n == 0) { - log_error("%s: no records found", name); + if (warn_missing) + log_error("%s: no records found", name); return -ESRCH; } @@ -618,7 +620,7 @@ static int resolve_rfc4501(sd_bus *bus, const char *name) { if (type == 0) type = arg_type ?: DNS_TYPE_A; - return resolve_record(bus, n, class, type); + return resolve_record(bus, n, class, type, true); invalid: log_error("Invalid DNS URI: %s", name); @@ -840,16 +842,34 @@ static int resolve_openpgp(sd_bus *bus, const char *address) { } domain++; - r = string_hashsum_sha224(address, domain - 1 - address, &hashed); + r = string_hashsum_sha256(address, domain - 1 - address, &hashed); if (r < 0) return log_error_errno(r, "Hashing failed: %m"); + strshorten(hashed, 56); + full = strjoina(hashed, "._openpgpkey.", domain); log_debug("Looking up \"%s\".", full); - return resolve_record(bus, full, - arg_class ?: DNS_CLASS_IN, - arg_type ?: DNS_TYPE_OPENPGPKEY); + r = resolve_record(bus, full, + arg_class ?: DNS_CLASS_IN, + arg_type ?: DNS_TYPE_OPENPGPKEY, false); + + if (IN_SET(r, -ENXIO, -ESRCH)) { /* NXDOMAIN or NODATA? */ + hashed = NULL; + r = string_hashsum_sha224(address, domain - 1 - address, &hashed); + if (r < 0) + return log_error_errno(r, "Hashing failed: %m"); + + full = strjoina(hashed, "._openpgpkey.", domain); + log_debug("Looking up \"%s\".", full); + + return resolve_record(bus, full, + arg_class ?: DNS_CLASS_IN, + arg_type ?: DNS_TYPE_OPENPGPKEY, true); + } + + return r; } static int resolve_tlsa(sd_bus *bus, const char *address) { @@ -881,7 +901,7 @@ static int resolve_tlsa(sd_bus *bus, const char *address) { return resolve_record(bus, full, arg_class ?: DNS_CLASS_IN, - arg_type ?: DNS_TYPE_TLSA); + arg_type ?: DNS_TYPE_TLSA, true); } static int show_statistics(sd_bus *bus) { @@ -1877,7 +1897,7 @@ int main(int argc, char **argv) { while (argv[optind]) { int k; - k = resolve_record(bus, argv[optind], arg_class, arg_type); + k = resolve_record(bus, argv[optind], arg_class, arg_type, true); if (r == 0) r = k; diff --git a/src/resolve/resolved-conf.c b/src/resolve/resolved-conf.c index 83f7596ac8..abf3263178 100644 --- a/src/resolve/resolved-conf.c +++ b/src/resolve/resolved-conf.c @@ -23,8 +23,19 @@ #include "extract-word.h" #include "parse-util.h" #include "resolved-conf.h" +#include "string-table.h" #include "string-util.h" +DEFINE_CONFIG_PARSE_ENUM(config_parse_dns_stub_listener_mode, dns_stub_listener_mode, DnsStubListenerMode, "Failed to parse DNS stub listener mode setting"); + +static const char* const dns_stub_listener_mode_table[_DNS_STUB_LISTENER_MODE_MAX] = { + [DNS_STUB_LISTENER_NO] = "no", + [DNS_STUB_LISTENER_UDP] = "udp", + [DNS_STUB_LISTENER_TCP] = "tcp", + [DNS_STUB_LISTENER_YES] = "yes", +}; +DEFINE_STRING_TABLE_LOOKUP_WITH_BOOLEAN(dns_stub_listener_mode, DnsStubListenerMode, DNS_STUB_LISTENER_YES); + int manager_add_dns_server_by_string(Manager *m, DnsServerType type, const char *word) { union in_addr_union address; int family, r, ifindex = 0; diff --git a/src/resolve/resolved-conf.h b/src/resolve/resolved-conf.h index e1fd2cceec..fc425a36b2 100644 --- a/src/resolve/resolved-conf.h +++ b/src/resolve/resolved-conf.h @@ -19,7 +19,19 @@ along with systemd; If not, see <http://www.gnu.org/licenses/>. ***/ +typedef enum DnsStubListenerMode DnsStubListenerMode; + +enum DnsStubListenerMode { + DNS_STUB_LISTENER_NO, + DNS_STUB_LISTENER_UDP, + DNS_STUB_LISTENER_TCP, + DNS_STUB_LISTENER_YES, + _DNS_STUB_LISTENER_MODE_MAX, + _DNS_STUB_LISTENER_MODE_INVALID = -1 +}; + #include "resolved-manager.h" +#include "resolved-dns-server.h" int manager_parse_config_file(Manager *m); @@ -33,4 +45,7 @@ const struct ConfigPerfItem* resolved_gperf_lookup(const char *key, unsigned len int config_parse_dns_servers(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata); int config_parse_search_domains(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata); -int config_parse_dnssec(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata); +int config_parse_dns_stub_listener_mode(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata); + +const char* dns_stub_listener_mode_to_string(DnsStubListenerMode p) _const_; +DnsStubListenerMode dns_stub_listener_mode_from_string(const char *s) _pure_; diff --git a/src/resolve/resolved-dns-scope.h b/src/resolve/resolved-dns-scope.h index 538bc61f81..01a83a76b2 100644 --- a/src/resolve/resolved-dns-scope.h +++ b/src/resolve/resolved-dns-scope.h @@ -26,7 +26,10 @@ typedef struct DnsScope DnsScope; #include "resolved-dns-cache.h" #include "resolved-dns-dnssec.h" #include "resolved-dns-packet.h" +#include "resolved-dns-query.h" +#include "resolved-dns-search-domain.h" #include "resolved-dns-server.h" +#include "resolved-dns-stream.h" #include "resolved-dns-zone.h" #include "resolved-link.h" diff --git a/src/resolve/resolved-dns-stream.h b/src/resolve/resolved-dns-stream.h index e6569678fa..4cdb4f6806 100644 --- a/src/resolve/resolved-dns-stream.h +++ b/src/resolve/resolved-dns-stream.h @@ -25,6 +25,7 @@ typedef struct DnsStream DnsStream; #include "resolved-dns-packet.h" #include "resolved-dns-transaction.h" +#include "resolved-manager.h" /* Streams are used by three subsystems: * diff --git a/src/resolve/resolved-dns-stub.c b/src/resolve/resolved-dns-stub.c index d263cedcd9..e76de6c06a 100644 --- a/src/resolve/resolved-dns-stub.c +++ b/src/resolve/resolved-dns-stub.c @@ -25,6 +25,9 @@ * IP and UDP header sizes */ #define ADVERTISE_DATAGRAM_SIZE_MAX (65536U-14U-20U-8U) +static int manager_dns_stub_udp_fd(Manager *m); +static int manager_dns_stub_tcp_fd(Manager *m); + static int dns_stub_make_reply_packet( uint16_t id, int rcode, @@ -354,66 +357,48 @@ static int on_dns_stub_packet(sd_event_source *s, int fd, uint32_t revents, void return 0; } -int manager_dns_stub_udp_fd(Manager *m) { +static int manager_dns_stub_udp_fd(Manager *m) { static const int one = 1; - union sockaddr_union sa = { .in.sin_family = AF_INET, .in.sin_port = htobe16(53), .in.sin_addr.s_addr = htobe32(INADDR_DNS_STUB), }; - + _cleanup_close_ int fd = -1; int r; if (m->dns_stub_udp_fd >= 0) return m->dns_stub_udp_fd; - m->dns_stub_udp_fd = socket(AF_INET, SOCK_DGRAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0); - if (m->dns_stub_udp_fd < 0) + fd = socket(AF_INET, SOCK_DGRAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0); + if (fd < 0) return -errno; - r = setsockopt(m->dns_stub_udp_fd, SOL_SOCKET, SO_REUSEADDR, &one, sizeof(one)); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &one, sizeof one) < 0) + return -errno; - r = setsockopt(m->dns_stub_udp_fd, IPPROTO_IP, IP_PKTINFO, &one, sizeof(one)); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, IPPROTO_IP, IP_PKTINFO, &one, sizeof one) < 0) + return -errno; - r = setsockopt(m->dns_stub_udp_fd, IPPROTO_IP, IP_RECVTTL, &one, sizeof(one)); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, IPPROTO_IP, IP_RECVTTL, &one, sizeof one) < 0) + return -errno; /* Make sure no traffic from outside the local host can leak to onto this socket */ - r = setsockopt(m->dns_stub_udp_fd, SOL_SOCKET, SO_BINDTODEVICE, "lo", 3); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, SOL_SOCKET, SO_BINDTODEVICE, "lo", 3) < 0) + return -errno; - r = bind(m->dns_stub_udp_fd, &sa.sa, sizeof(sa.in)); - if (r < 0) { - r = -errno; - goto fail; - } + if (bind(fd, &sa.sa, sizeof(sa.in)) < 0) + return -errno; - r = sd_event_add_io(m->event, &m->dns_stub_udp_event_source, m->dns_stub_udp_fd, EPOLLIN, on_dns_stub_packet, m); + r = sd_event_add_io(m->event, &m->dns_stub_udp_event_source, fd, EPOLLIN, on_dns_stub_packet, m); if (r < 0) - goto fail; + return r; (void) sd_event_source_set_description(m->dns_stub_udp_event_source, "dns-stub-udp"); + m->dns_stub_udp_fd = fd; + fd = -1; return m->dns_stub_udp_fd; - -fail: - m->dns_stub_udp_fd = safe_close(m->dns_stub_udp_fd); - return r; } static int on_dns_stub_stream_packet(DnsStream *s) { @@ -461,102 +446,83 @@ static int on_dns_stub_stream(sd_event_source *s, int fd, uint32_t revents, void return 0; } -int manager_dns_stub_tcp_fd(Manager *m) { +static int manager_dns_stub_tcp_fd(Manager *m) { static const int one = 1; - union sockaddr_union sa = { .in.sin_family = AF_INET, .in.sin_addr.s_addr = htobe32(INADDR_DNS_STUB), .in.sin_port = htobe16(53), }; - + _cleanup_close_ int fd = -1; int r; if (m->dns_stub_tcp_fd >= 0) return m->dns_stub_tcp_fd; - m->dns_stub_tcp_fd = socket(AF_INET, SOCK_STREAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0); - if (m->dns_stub_tcp_fd < 0) + fd = socket(AF_INET, SOCK_STREAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0); + if (fd < 0) return -errno; - r = setsockopt(m->dns_stub_tcp_fd, IPPROTO_IP, IP_TTL, &one, sizeof(one)); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, IPPROTO_IP, IP_TTL, &one, sizeof one) < 0) + return -errno; - r = setsockopt(m->dns_stub_tcp_fd, SOL_SOCKET, SO_REUSEADDR, &one, sizeof(one)); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &one, sizeof one) < 0) + return -errno; - r = setsockopt(m->dns_stub_tcp_fd, IPPROTO_IP, IP_PKTINFO, &one, sizeof(one)); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, IPPROTO_IP, IP_PKTINFO, &one, sizeof one) < 0) + return -errno; - r = setsockopt(m->dns_stub_tcp_fd, IPPROTO_IP, IP_RECVTTL, &one, sizeof(one)); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, IPPROTO_IP, IP_RECVTTL, &one, sizeof one) < 0) + return -errno; /* Make sure no traffic from outside the local host can leak to onto this socket */ - r = setsockopt(m->dns_stub_tcp_fd, SOL_SOCKET, SO_BINDTODEVICE, "lo", 3); - if (r < 0) { - r = -errno; - goto fail; - } + if (setsockopt(fd, SOL_SOCKET, SO_BINDTODEVICE, "lo", 3) < 0) + return -errno; - r = bind(m->dns_stub_tcp_fd, &sa.sa, sizeof(sa.in)); - if (r < 0) { - r = -errno; - goto fail; - } + if (bind(fd, &sa.sa, sizeof(sa.in)) < 0) + return -errno; - r = listen(m->dns_stub_tcp_fd, SOMAXCONN); - if (r < 0) { - r = -errno; - goto fail; - } + if (listen(fd, SOMAXCONN) < 0) + return -errno; - r = sd_event_add_io(m->event, &m->dns_stub_tcp_event_source, m->dns_stub_tcp_fd, EPOLLIN, on_dns_stub_stream, m); + r = sd_event_add_io(m->event, &m->dns_stub_tcp_event_source, fd, EPOLLIN, on_dns_stub_stream, m); if (r < 0) - goto fail; + return r; (void) sd_event_source_set_description(m->dns_stub_tcp_event_source, "dns-stub-tcp"); + m->dns_stub_tcp_fd = fd; + fd = -1; return m->dns_stub_tcp_fd; - -fail: - m->dns_stub_tcp_fd = safe_close(m->dns_stub_tcp_fd); - return r; } int manager_dns_stub_start(Manager *m) { - int r; + const char *t = "UDP"; + int r = 0; assert(m); - r = manager_dns_stub_udp_fd(m); - if (r == -EADDRINUSE) - goto eaddrinuse; - if (r < 0) - return r; - - r = manager_dns_stub_tcp_fd(m); - if (r == -EADDRINUSE) - goto eaddrinuse; - if (r < 0) - return r; + if (IN_SET(m->dns_stub_listener_mode, DNS_STUB_LISTENER_YES, DNS_STUB_LISTENER_UDP)) + r = manager_dns_stub_udp_fd(m); - return 0; + if (r >= 0 && + IN_SET(m->dns_stub_listener_mode, DNS_STUB_LISTENER_YES, DNS_STUB_LISTENER_TCP)) { + t = "TCP"; + r = manager_dns_stub_tcp_fd(m); + } -eaddrinuse: - log_warning("Another process is already listening on 127.0.0.53:53. Turning off local DNS stub support."); - manager_dns_stub_stop(m); + if (IN_SET(r, -EADDRINUSE, -EPERM)) { + if (r == -EADDRINUSE) + log_warning_errno(r, + "Another process is already listening on %s socket 127.0.0.53:53.\n" + "Turning off local DNS stub support.", t); + else + log_warning_errno(r, + "Failed to listen on %s socket 127.0.0.53:53: %m.\n" + "Turning off local DNS stub support.", t); + manager_dns_stub_stop(m); + } else if (r < 0) + return log_error_errno(r, "Failed to listen on %s socket 127.0.0.53:53: %m", t); return 0; } diff --git a/src/resolve/resolved-dns-stub.h b/src/resolve/resolved-dns-stub.h index fce4d25ede..12b86f6753 100644 --- a/src/resolve/resolved-dns-stub.h +++ b/src/resolve/resolved-dns-stub.h @@ -24,8 +24,5 @@ /* 127.0.0.53 in native endian */ #define INADDR_DNS_STUB ((in_addr_t) 0x7f000035U) -int manager_dns_stub_udp_fd(Manager *m); -int manager_dns_stub_tcp_fd(Manager *m); - void manager_dns_stub_stop(Manager *m); int manager_dns_stub_start(Manager *m); diff --git a/src/resolve/resolved-dns-transaction.h b/src/resolve/resolved-dns-transaction.h index 96b066845d..5a1df70422 100644 --- a/src/resolve/resolved-dns-transaction.h +++ b/src/resolve/resolved-dns-transaction.h @@ -59,6 +59,8 @@ enum DnsTransactionSource { #include "resolved-dns-packet.h" #include "resolved-dns-question.h" #include "resolved-dns-scope.h" +#include "resolved-dns-server.h" +#include "resolved-dns-stream.h" struct DnsTransaction { DnsScope *scope; diff --git a/src/resolve/resolved-gperf.gperf b/src/resolve/resolved-gperf.gperf index 2fd56bce26..446f85cdf4 100644 --- a/src/resolve/resolved-gperf.gperf +++ b/src/resolve/resolved-gperf.gperf @@ -14,9 +14,10 @@ struct ConfigPerfItem; %struct-type %includes %% -Resolve.DNS, config_parse_dns_servers, DNS_SERVER_SYSTEM, 0 -Resolve.FallbackDNS, config_parse_dns_servers, DNS_SERVER_FALLBACK, 0 -Resolve.Domains, config_parse_search_domains, 0, 0 -Resolve.LLMNR, config_parse_resolve_support, 0, offsetof(Manager, llmnr_support) -Resolve.DNSSEC, config_parse_dnssec_mode, 0, offsetof(Manager, dnssec_mode) -Resolve.Cache, config_parse_bool, 0, offsetof(Manager, enable_cache) +Resolve.DNS, config_parse_dns_servers, DNS_SERVER_SYSTEM, 0 +Resolve.FallbackDNS, config_parse_dns_servers, DNS_SERVER_FALLBACK, 0 +Resolve.Domains, config_parse_search_domains, 0, 0 +Resolve.LLMNR, config_parse_resolve_support, 0, offsetof(Manager, llmnr_support) +Resolve.DNSSEC, config_parse_dnssec_mode, 0, offsetof(Manager, dnssec_mode) +Resolve.Cache, config_parse_bool, 0, offsetof(Manager, enable_cache) +Resolve.DNSStubListener, config_parse_dns_stub_listener_mode, 0, offsetof(Manager, dns_stub_listener_mode) diff --git a/src/resolve/resolved-link.h b/src/resolve/resolved-link.h index 6a2343f9f7..c9b2a58c34 100644 --- a/src/resolve/resolved-link.h +++ b/src/resolve/resolved-link.h @@ -29,6 +29,7 @@ typedef struct Link Link; typedef struct LinkAddress LinkAddress; #include "resolved-dns-rr.h" +#include "resolved-dns-scope.h" #include "resolved-dns-search-domain.h" #include "resolved-dns-server.h" #include "resolved-manager.h" diff --git a/src/resolve/resolved-manager.c b/src/resolve/resolved-manager.c index 92ade820ac..40f08e8044 100644 --- a/src/resolve/resolved-manager.c +++ b/src/resolve/resolved-manager.c @@ -501,6 +501,7 @@ int manager_new(Manager **ret) { m->mdns_support = RESOLVE_SUPPORT_NO; m->dnssec_mode = DEFAULT_DNSSEC_MODE; m->enable_cache = true; + m->dns_stub_listener_mode = DNS_STUB_LISTENER_UDP; m->read_resolv_conf = true; m->need_builtin_fallbacks = true; m->etc_hosts_last = m->etc_hosts_mtime = USEC_INFINITY; diff --git a/src/resolve/resolved-manager.h b/src/resolve/resolved-manager.h index deebd8e484..6b2208ed94 100644 --- a/src/resolve/resolved-manager.h +++ b/src/resolve/resolved-manager.h @@ -30,6 +30,7 @@ typedef struct Manager Manager; +#include "resolved-conf.h" #include "resolved-dns-query.h" #include "resolved-dns-search-domain.h" #include "resolved-dns-server.h" @@ -47,6 +48,7 @@ struct Manager { ResolveSupport mdns_support; DnssecMode dnssec_mode; bool enable_cache; + DnsStubListenerMode dns_stub_listener_mode; /* Network */ Hashmap *links; diff --git a/src/resolve/resolved.conf.in b/src/resolve/resolved.conf.in index 3bd8389c88..60afa151e3 100644 --- a/src/resolve/resolved.conf.in +++ b/src/resolve/resolved.conf.in @@ -18,3 +18,4 @@ #LLMNR=yes #DNSSEC=@DEFAULT_DNSSEC_MODE@ #Cache=yes +#DNSStubListener=udp diff --git a/src/shared/gcrypt-util.h b/src/shared/gcrypt-util.h index cf33b3c59c..1da12a32be 100644 --- a/src/shared/gcrypt-util.h +++ b/src/shared/gcrypt-util.h @@ -37,3 +37,11 @@ static inline int string_hashsum_sha224(const char *s, size_t len, char **out) { return -EOPNOTSUPP; #endif } + +static inline int string_hashsum_sha256(const char *s, size_t len, char **out) { +#ifdef HAVE_GCRYPT + return string_hashsum(s, len, GCRY_MD_SHA256, out); +#else + return -EOPNOTSUPP; +#endif +} diff --git a/src/shared/install.c b/src/shared/install.c index 5f12fb447f..60a6d1312d 100644 --- a/src/shared/install.c +++ b/src/shared/install.c @@ -1020,7 +1020,7 @@ static int config_parse_alias( type = unit_name_to_type(name); if (!unit_type_may_alias(type)) return log_syntax(unit, LOG_WARNING, filename, line, 0, - "Aliases are not allowed for %s units, ignoring.", + "Alias= is not allowed for %s units, ignoring.", unit_type_to_string(type)); return config_parse_strv(unit, filename, line, section, section_line, @@ -1098,7 +1098,7 @@ static int config_parse_default_instance( return 0; if (!unit_name_is_valid(name, UNIT_NAME_TEMPLATE)) return log_syntax(unit, LOG_WARNING, filename, line, 0, - "DefaultInstance only makes sense for template units, ignoring."); + "DefaultInstance= only makes sense for template units, ignoring."); r = install_full_printf(i, rvalue, &printed); if (r < 0) diff --git a/src/systemctl/systemctl.c b/src/systemctl/systemctl.c index bb6002e8ef..18a8a4f248 100644 --- a/src/systemctl/systemctl.c +++ b/src/systemctl/systemctl.c @@ -3936,7 +3936,7 @@ static void print_status_info( argv = strv_join(p->argv, " "); printf(" Process: "PID_FMT" %s=%s ", p->pid, p->name, strna(argv)); - good = is_clean_exit_lsb(p->code, p->status, NULL); + good = is_clean_exit(p->code, p->status, EXIT_CLEAN_DAEMON, NULL); if (!good) { on = ansi_highlight_red(); off = ansi_normal(); diff --git a/src/sysv-generator/sysv-generator.c b/src/sysv-generator/sysv-generator.c index 39821687b9..c2c80175a2 100644 --- a/src/sysv-generator/sysv-generator.c +++ b/src/sysv-generator/sysv-generator.c @@ -25,6 +25,7 @@ #include "alloc-util.h" #include "dirent-util.h" +#include "exit-status.h" #include "fd-util.h" #include "fileio.h" #include "hashmap.h" @@ -199,6 +200,13 @@ static int generate_unit_file(SysvStub *s) { if (s->pid_file) fprintf(f, "PIDFile=%s\n", s->pid_file); + /* Consider two special LSB exit codes a clean exit */ + if (s->has_lsb) + fprintf(f, + "SuccessExitStatus=%i %i\n", + EXIT_NOTINSTALLED, + EXIT_NOTCONFIGURED); + fprintf(f, "ExecStart=%s start\n" "ExecStop=%s stop\n", diff --git a/src/test/test-path-util.c b/src/test/test-path-util.c index 164a10d8a8..0b10d8e25e 100644 --- a/src/test/test-path-util.c +++ b/src/test/test-path-util.c @@ -511,7 +511,24 @@ static void test_hidden_or_backup_file(void) { assert_se(!hidden_or_backup_file("test.dpkg-old.foo")); } +static void test_systemd_installation_has_version(const char *path) { + int r; + const unsigned versions[] = {0, 231, atoi(PACKAGE_VERSION), 999}; + unsigned i; + + for (i = 0; i < ELEMENTSOF(versions); i++) { + r = systemd_installation_has_version(path, versions[i]); + assert_se(r >= 0); + log_info("%s has systemd >= %u: %s", + path ?: "Current installation", versions[i], yes_no(r)); + } +} + int main(int argc, char **argv) { + log_set_max_level(LOG_DEBUG); + log_parse_environment(); + log_open(); + test_path(); test_find_binary(argv[0]); test_prefixes(); @@ -526,5 +543,7 @@ int main(int argc, char **argv) { test_filename_is_valid(); test_hidden_or_backup_file(); + test_systemd_installation_has_version(argv[1]); /* NULL is OK */ + return 0; } diff --git a/src/tty-ask-password-agent/tty-ask-password-agent.c b/src/tty-ask-password-agent/tty-ask-password-agent.c index 8851af449d..b45490be1a 100644 --- a/src/tty-ask-password-agent/tty-ask-password-agent.c +++ b/src/tty-ask-password-agent/tty-ask-password-agent.c @@ -827,7 +827,7 @@ static int ask_on_consoles(int argc, char *argv[]) { break; } - if (!is_clean_exit(status.si_code, status.si_status, NULL)) + if (!is_clean_exit(status.si_code, status.si_status, EXIT_CLEAN_DAEMON, NULL)) log_error("Password agent failed with: %d", status.si_status); terminate_agents(pids); diff --git a/src/udev/udevadm-control.c b/src/udev/udevadm-control.c index 989decbe95..6f8e96a123 100644 --- a/src/udev/udevadm-control.c +++ b/src/udev/udevadm-control.c @@ -20,6 +20,7 @@ #include <string.h> #include <unistd.h> +#include "time-util.h" #include "udev-util.h" #include "udev.h" @@ -60,7 +61,7 @@ static int adm_control(struct udev *udev, int argc, char *argv[]) { }; if (getuid() != 0) { - fprintf(stderr, "root privileges required\n"); + log_error("root privileges required"); return 1; } @@ -81,7 +82,7 @@ static int adm_control(struct udev *udev, int argc, char *argv[]) { i = util_log_priority(optarg); if (i < 0) { - fprintf(stderr, "invalid number '%s'\n", optarg); + log_error("invalid number '%s'", optarg); return rc; } if (udev_ctrl_send_set_log_level(uctrl, util_log_priority(optarg), timeout) < 0) @@ -110,7 +111,7 @@ static int adm_control(struct udev *udev, int argc, char *argv[]) { break; case 'p': if (strchr(optarg, '=') == NULL) { - fprintf(stderr, "expect <KEY>=<value> instead of '%s'\n", optarg); + log_error("expect <KEY>=<value> instead of '%s'", optarg); return rc; } if (udev_ctrl_send_set_env(uctrl, optarg, timeout) < 0) @@ -124,7 +125,7 @@ static int adm_control(struct udev *udev, int argc, char *argv[]) { i = strtoul(optarg, &endp, 0); if (endp[0] != '\0' || i < 1) { - fprintf(stderr, "invalid number '%s'\n", optarg); + log_error("invalid number '%s'", optarg); return rc; } if (udev_ctrl_send_set_children_max(uctrl, i, timeout) < 0) @@ -134,13 +135,21 @@ static int adm_control(struct udev *udev, int argc, char *argv[]) { break; } case 't': { + usec_t s; int seconds; + int r; - seconds = atoi(optarg); - if (seconds >= 0) + r = parse_sec(optarg, &s); + if (r < 0) + return log_error_errno(r, "Failed to parse timeout value '%s'.", optarg); + + if (((s + USEC_PER_SEC - 1) / USEC_PER_SEC) > INT_MAX) + log_error("Timeout value is out of range."); + else { + seconds = s != USEC_INFINITY ? (int) ((s + USEC_PER_SEC - 1) / USEC_PER_SEC) : INT_MAX; timeout = seconds; - else - fprintf(stderr, "invalid timeout value\n"); + rc = 0; + } break; } case 'h': @@ -150,9 +159,9 @@ static int adm_control(struct udev *udev, int argc, char *argv[]) { } if (optind < argc) - fprintf(stderr, "Extraneous argument: %s\n", argv[optind]); + log_error("Extraneous argument: %s", argv[optind]); else if (optind == 1) - fprintf(stderr, "Option missing\n"); + log_error("Option missing"); return rc; } diff --git a/src/udev/udevd.c b/src/udev/udevd.c index 19f1c29198..535d317c27 100644 --- a/src/udev/udevd.c +++ b/src/udev/udevd.c @@ -1738,8 +1738,13 @@ int main(int argc, char *argv[]) { log_info("starting version " VERSION); /* connect /dev/null to stdin, stdout, stderr */ - if (log_get_max_level() < LOG_DEBUG) - (void) make_null_stdio(); + if (log_get_max_level() < LOG_DEBUG) { + r = make_null_stdio(); + if (r < 0) + log_warning_errno(r, "Failed to redirect standard streams to /dev/null: %m"); + } + + pid = fork(); switch (pid) { diff --git a/system-preset/90-systemd.preset b/system-preset/90-systemd.preset index 0f494b7552..6b5349dc8d 100644 --- a/system-preset/90-systemd.preset +++ b/system-preset/90-systemd.preset @@ -18,7 +18,6 @@ enable systemd-resolved.service enable systemd-networkd-wait-online.service disable console-getty.service -disable console-shell.service disable debug-shell.service disable halt.target diff --git a/units/.gitignore b/units/.gitignore index 47e99154ee..8f4949258e 100644 --- a/units/.gitignore +++ b/units/.gitignore @@ -1,8 +1,6 @@ /user@.service.m4 /console-getty.service /console-getty.service.m4 -/console-shell.service -/console-shell.service.m4 /container-getty@.service /container-getty@.service.m4 /debug-shell.service diff --git a/units/console-shell.service.m4.in b/units/console-shell.service.m4.in deleted file mode 100644 index a345ec25d4..0000000000 --- a/units/console-shell.service.m4.in +++ /dev/null @@ -1,31 +0,0 @@ -# This file is part of systemd. -# -# systemd is free software; you can redistribute it and/or modify it -# under the terms of the GNU Lesser General Public License as published by -# the Free Software Foundation; either version 2.1 of the License, or -# (at your option) any later version. - -[Unit] -Description=Console Shell -Documentation=man:sulogin(8) -After=systemd-user-sessions.service plymouth-quit-wait.service -m4_ifdef(`HAVE_SYSV_COMPAT', -After=rc-local.service -)m4_dnl -Before=getty.target - -[Service] -Environment=HOME=/root -WorkingDirectory=-/root -ExecStart=-@SULOGIN@ -ExecStopPost=-@SYSTEMCTL@ poweroff -Type=idle -StandardInput=tty-force -StandardOutput=inherit -StandardError=inherit -KillMode=process -IgnoreSIGPIPE=no -SendSIGHUP=yes - -[Install] -WantedBy=getty.target diff --git a/units/initrd-switch-root.target b/units/initrd-switch-root.target index f34768790b..934d82f667 100644 --- a/units/initrd-switch-root.target +++ b/units/initrd-switch-root.target @@ -12,5 +12,5 @@ DefaultDependencies=no Requires=initrd-switch-root.service Before=initrd-switch-root.service AllowIsolate=yes -Wants=initrd-udevadm-cleanup-db.service initrd-root-fs.target initrd-fs.target systemd-journald.service +Wants=initrd-udevadm-cleanup-db.service initrd-root-fs.target initrd-fs.target systemd-journald.service initrd-cleanup.service After=initrd-udevadm-cleanup-db.service initrd-root-fs.target initrd-fs.target emergency.service emergency.target |