Currently, 'make distcheck' fails with an error such as this:
srcdir=../../po /usr/bin/intltool-update -m
The following files contain translations and are currently not in use.
Please consider adding these to the POTFILES.in file, located in the po/ directory.
build2/src/core/org.freedesktop.systemd1.policy.in
build3/src/core/org.freedesktop.systemd1.policy.in
[...]
This is caused by a new behavior of autmake 1.15 which changed the
location of the build tree during 'make distcheck', and the fact that
intltool doesn't yet ignore that paths.
We used to have a workaround in configure.ac that makes the failing call
a no-op, but it was accidentially removed in 23756070
("remove gudev and gtk-doc").
Bring back that snipet for now, until intltool and automake sorted out
their issues and like each other again.
Also see
https://bugs.launchpad.net/intltool/+bug/1117944
When calling the build commands from another directory than the toplevel:
mkdir build
cd build
../configure [...]
... the compilation fails with the following error:
GEN man/systemd.directives.xml
Traceback (most recent call last):
File "../tools/make-directive-index.py", line 313, in make_page
_extract_directives(directive_groups, formatting, page)
File "../tools/make-directive-index.py", line 191, in _extract_directives
t = xml_parse(page)
File "/home/martin/upstream/systemd/tools/xml_helper.py", line 30, in xml_parse
doc = tree.parse(page, _parser)
File "lxml.etree.pyx", line 3301, in lxml.etree.parse (src/lxml/lxml.etree.c:72453)
File "parser.pxi", line 1791, in lxml.etree._parseDocument (src/lxml/lxml.etree.c:105915)
File "parser.pxi", line 1817, in lxml.etree._parseDocumentFromURL (src/lxml/lxml.etree.c:106214)
File "parser.pxi", line 1721, in lxml.etree._parseDocFromFile (src/lxml/lxml.etree.c:105213)
File "parser.pxi", line 1122, in lxml.etree._BaseParser._parseDocFromFile (src/lxml/lxml.etree.c:100163)
File "parser.pxi", line 580, in lxml.etree._ParserContext._handleParseResultDoc (src/lxml/lxml.etree.c:94286)
File "parser.pxi", line 690, in lxml.etree._handleParseResult (src/lxml/lxml.etree.c:95722)
File "parser.pxi", line 618, in lxml.etree._raiseParseError (src/lxml/lxml.etree.c:94754)
OSError: Error reading file 'man/bootup.xml': failed to load external entity "man/bootup.xml"
That is because the file names in the XML_FILES array are not relative
to $(top_srcdir), and hence ../tools/make-directive-index.py is called
with non-existant arguments.
To fix this, call patsubst when generating SOURCE_XML_FILES from
NON_INDEX_XML_FILES.
Commit 72c0a2c25 ("everywhere: port everything to sigprocmask_many()
and friends") reworked code tree-wide to use the new sigprocmask_many()
helper. In this, it caused a regression in pam_setup, because it
dropped a line to initialize the 'ss' signal mask which is later used
in sigwait().
While at it, move the variable declaration to an inner scope.
-ENOSYS is returned from kmod_module_probe_insert_module() if a module isn't
available, not -ENOENT. Don't spit out a warning in that case unless the
warn_if_unavailable flag is set.
Also factor out the condition into an own variable for better readability.
Currently, the following command sequence fails:
make distclean
./autogen.sh c
make distcheck
That's because the command invoked to build man/systemd.directives.xml needs
man/custom-entities.ent to function, which itself isn't a dependency.
The $(filter-out $<,$^) logic used to filter out everything from the
prerequisites except for the first word, which doesn't work anymore
now. Use $(SOURCE_XML_FILES) instead.
Make the build sys error out on missing function prototypes, missing
variable declarations, implicit function declarations or forgotten return
statements.
None of these conditions are acceptable, and by making them hard errors, the
build bots can detect them earlier.
Traditionally, we used to warn about ipv6 being a module or being
unavailable. This was changed in b4aa82f16 ("kmod-setup: don't warn
when ipv6 can't be loaded") in a way that neither of the two conditions
will cause a log message.
Now, while running a setup without any IPv6 is completely fine and
shouldn't cause any warning, we should still warn about ipv6 being a
module instead of built-in.
To achieve this, split the boolean warn flag into two: one for a
feature not being built-in but shipped as a module, and one to
print an error when a module is entirely unavailable.
We will, however, still warn if kmod returns anything else than
-ENOENT in the attempt of loading the module, and at the very least,
turn the message into a debug log.
A README.md allows us to include a badge for the current build status of
Semaphore CI. Keep other information in this file minimal, and avoid
information duplication but point readers to the official README.
Skip /proc/mountinfo entries for which libmount returns a NULL pointer
for 'source' or 'target'. This happened on Semaphore CI's build servers
when the test suite is run.
Fix a regression caused by 4034a06d ("util: rework word parsing and c
unescaping code") which broke octal escape sequences.
The reason for this breakage is that cunescape_one() expects 4 characters
in an octal encoding, which is a stray left-over from the old code which
operated on different variables to make the length check.
While at it, add a test case to prevent the same thing from happening
again.
The cunescape() helper function used to handle unknown escaping sequences
gracefully by copying them over verbatim.
Commit 527b7a42 ("util: rework cunescape(), improve error handling") added
a flag to make that behavior optional, and changed to default to error out
with -EINVAL otherwise.
However, config_parse_exec(), which is used to parse the
Exec{Start,Stop}{Post,Pre,} directives of unit files, was not changed along
with that commit, which means that directives with improperly escaped
command line strings are no longer parsed.
Relevant bugreports include:
https://bugs.freedesktop.org/show_bug.cgi?id=90794https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=787256
Fix this by passing UNESCAPE_RELAX to config_parse_exec() in order to
restore the original behavior.
This reverts commit 6096d9cc. As discussed on the mailing list, we
should accept some formal incorrectness in the dependency here, and
not rebuild the man pages every time Makefile.am changes - xsltproc
is simply too expensive.
Instead, let's move man/custom-entities.ent from DISTCLEANFILES to
CLEANFILES, so a 'make clean' is sufficient to actually make changes
in Makefile.am efficient for the contents of the man pages.
When Makefile.am is modified, make sure custom-entities.ent is rebuilt.
After all, $(substitutions) is defined there, so changes of that variable
must be reflected in the resulting file.
Not that all functionality has been ported over to logind, the old
implementation can be removed. There goes one of the oldest parts of
the systemd code base.
Drop the code which communicates with shutdownd via its private socket,
and use the functionality in logind instead.
The code pathes which talk to logind have to create their own ad-hoc
bus connection because by default, systemctl connects to systemd's
private socket.
Port over more code from shutdownd and teach logind to write /run/nologin at
least 5 minutes before the system is going down, and
/run/systemd/shutdown/scheduled when a shutdown is scheduled.
Add a timer to print UTMP wall messages so that it repeatedly informs users
about a scheduled shutdown:
* every 1 minute with less than 10 minutes to go
* every 15 minutes with less than 60 minutes to go
* every 30 minutes with less than 180 minutes (3 hours) to go
* every 60 minutes if more than that to go
This functionality only active if the .EnableWallMessages DBus property
is set to true. Also, a custom string can be added to the wall message,
set through the WallMessagePrefix property.
Instead of looking up the tty from STDIN, let utmp_wall() take an argument
to specify an origin tty for the wall message. Only if that argument is
NULL do the STDIN lookup.
Also add an void *userdata argument that is handed back to the callback
function.
Add a method called ScheduleShutdown in org.freedesktop.login1.Manager
which adds a timer to shut down the system at a later point in time.
The first argument holds the type of the schedule that is about to
happen, and must be one of 'reboot', 'halt' or 'poweroff'.
The second argument specifies the absolute time, based on
CLOCK_REALTIME in nanoseconds, at which the the operation should be
executed.
To cancel a previously scheduled shutdown, the CancelScheduledShutdown()
can be called, which returns a bool, indicating whether a scheduled
timeout was cancelled.
Also add a new property called ScheduledShutdown which returns the
equivalent to what was passed in via ScheduleShutdown, as '(st)' type.
Factor out the code to ask polkit for authorization from
method_do_shutdown_or_sleep() into an own function called
verify_shutdown_creds().
This is needed in order to also use the same checks when shutdown
operations are scheduled. For that, it's also necessary to allow
NULL values for that action{,_multiple_sessions,_ignore_inhibit)
arguments, which will suppress the call if no action string is
passed.
Instead of open-coding the delayed action and inhibit timeout logic,
switch over to a real sd_event_source based implementation.
This is not only easier to read but also allows us to add more timers
in the future.
Fixes:
CC src/core/libsystemd_core_la-device.lo
src/core/device.c: In function 'device_serialize':
src/core/device.c:169:1: warning: control reaches end of non-void function [-Wreturn-type]
}
^
Commit e792e890f ("path-util: don't eat up ENOENT in
path_is_mount_point()") changed path_is_mount_point() so it doesn't hide
-ENOENT from its caller. This causes all boots to fail early in case
any of the mount points does not exist (for instance, when kdbus isn't
loaded, /sys/fs/kdbus is missing).
Fix this by returning 0 from mount_one() if path_is_mount_point()
returned -ENOENT.
* kill unnecessary {}
* add newlines where appropriate
* remove dead code
* reorder variable declarations
* fix more return code logic
* pass O_CLOEXEC to all open*() calles
* use safe_close() where possible
Retrieve the handle to procfs in main(), and pass it functions
that need it. Kill the global variables.
Also, refactor lots of code in svg_title(). There's no need to access any
global variables from there either, and we really should return proper
errors from there as well.
Don't blindly exit() from random functions, but return a proper error
and upchain error conditions.
squash! bootchart: clean up control flow logic
When pread() returns "0", it's a read failure, so don't make the caller think
log_sample() was successful, return meaningful error code instead of 0.
After some reconsideration, we decided to move the binary protocol
back to 64-bit wide UIDs and GIDs. After all, it should be possible
to redefine [gu]id_t to uint64_t and things should continue to
work. As we want to avoid such data types in kdbus.h, let's move
back to 64-bit values and be safe.
In sd-bus, we have to do a translation between uint64_t and gid_t
now for supplementary gids.
Some inline comments have also been updated in kdbus upstream.
Fixes the warning below.
src/shared/condition.c: In function ‘condition_new’:
src/shared/condition.c:47:27: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses]
assert(!parameter == (type == CONDITION_NULL));
^
src/shared/macro.h:42:44: note: in definition of macro ‘_unlikely_’
#define _unlikely_(x) (__builtin_expect(!!(x),0))
^
src/shared/macro.h:226:22: note: in expansion of macro ‘assert_se’
#define assert(expr) assert_se(expr)
^
src/shared/condition.c:47:9: note: in expansion of macro ‘assert’
assert(!parameter == (type == CONDITION_NULL));
^
gcc5 spits out a warning about test-hashmap.c:
CC src/test/test-hashmap.o
src/test/test-hashmap.c: In function ‘test_string_compare_func’:
src/test/test-hashmap.c:76:79: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses]
Fix the following gcc5 warning:
CC src/shared/libsystemd_shared_la-unit-name.lo
src/shared/unit-name.c: In function 'unit_name_is_valid':
src/shared/unit-name.c:102:34: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses]
if (!template_ok == TEMPLATE_VALID && at+1 == e)
^
The kernel part of kdbus does not allow userspace to make any assumption on
which unique address the first user on the bus will get.
Instead, use sd_bus_get_unique_name() to get the server's address.
CC src/shared/libsystemd_shared_la-time-util.lo
src/shared/time-util.c: In function 'parse_nsec':
src/shared/time-util.c:789:25: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses]
if (!*s != 0)
^
Follow two small changes in the kdbus API:
* Flags are now returned in cmd->return_flags by KDBUS_CMD_NAME_ACQUIRE
* struct kdbus_item_list has been dropped. The information stored in
this struct was redundant since awhile already, as all commands
report their returned slice size anyway.
Include <sys/socket.h> from util.h and bus-message.h in order to
build errors like the ones below on kdbus enabled systems.
./src/shared/util.h:976:32: warning: its scope is only this definition or declaration, which is probably not what you want
In file included from src/libsystemd/sd-bus/bus-kernel.c:37:0:
./src/shared/util.h:1081:28: warning: 'struct msghdr' declared inside parameter list
void cmsg_close_all(struct msghdr *mh);
^
CC src/libsystemd/sd-bus/libsystemd_la-bus-creds.lo
In file included from src/libsystemd/sd-bus/bus-creds.c:25:0:
./src/shared/util.h:976:32: warning: 'struct ucred' declared inside parameter list
int getpeercred(int fd, struct ucred *ucred);
^
strempty() will return an empty string in case the input parameter is
a NULL pointer. The correct test to check for an empty string is
isempty(), so use that instead.
This fixes a regression from commit 17a1c59 ("core/mount: filter out
noauto,auto,nofail,fail options").
Catch up with latest changes in kdbus.ko:
* Signals can be sent as unicast now, hence they need to be marked as
such with the KDBUS_MSG_SIGNAL in the message flags.
* Follow ioctl number change for KDBUS_CMD_FREE
sd_bus_creds_get_well_known_names() fails with -ENODATA in case the
message has no names attached, which is intended behavior if the
remote connection didn't own any names at the time of sending.
The function already deals with 'sender_names' being an empty strv,
so we can just continue in such cases.
Messages to destinations that are not currently owned by any bus connection
will cause kdbus related function to return with either -ENXIO or -ESRCH.
Such conditions should not make the proxyd terminate but send a sane
SD_BUS_ERROR_NAME_HAS_NO_OWNER error reply to the proxied connection.
Sync kdbus.h with upstream changes:
* Two optional cancellation points where added for synchronously
blocking KDBUS_CMD_SEND commands: A sigmask to change the mask
of accepted signals before the task is put to sleep, and a
generic file descriptor that can be written to, in order to cancel
the command. Both methods are currently unused.
* The KDBUS_CMD_CANCEL ioctl was removed. sd-bus was never using
that command, so there's no change needed.
* Some kerneldoc fixes
Implement a recent change in the kdbus pool logic:
PAYLOAD_VEC_OFF items are now referencing offsets relative to the
connection's pool, not to the item itself. Follow this change in
sd-bus.
Turns out we can just do kmod_setup() earlier, before we do mount_setup(),
so there's no need for mount_setup_late() anymore. Instead, put kdbusfs in
mount_table[].
kdbus has seen a larger update than expected lately, most notably with
kdbusfs, a file system to expose the kdbus control files:
* Each time a file system of this type is mounted, a new kdbus
domain is created.
* The layout inside each mount point is the same as before, except
that domains are not hierarchically nested anymore.
* Domains are therefore also unnamed now.
* Unmounting a kdbusfs will automatically also detroy the
associated domain.
* Hence, the action of creating a kdbus domain is now as
privileged as mounting a filesystem.
* This way, we can get around creating dev nodes for everything,
which is last but not least something that is not limited by
20-bit minor numbers.
The kdbus specific bits in nspawn have all been dropped now, as nspawn
can rely on the container OS to set up its own kdbus domain, simply by
mounting a new instance.
A new set of mounts has been added to mount things *after* the kernel
modules have been loaded. For now, only kdbus is in this set, which is
invoked with mount_setup_late().
There are issues to investigate on with policies shipped by some
packages, which we'll address later. Move that topic out of the
way for now to bring sd-bus in sync with upstream kdbus.
kdbus learned parsing the attach flags for the KDBUS_CMD_BUS_CREATOR_INFO
ioctl. Bits not set in this mask will not be exported. Set that field to
_KDBUS_ATTACH_ALL for now.
Signed-off-by: Daniel Mack <daniel@zonque.org>
In order to check for matching policy entries at message transfers, we
have to consider the following:
* check the currently owned names of both the sending and the receiving
peer. If the sending peer is connected via kdbus, the currently owned
names are already attached to the message. If it was originated by the
connection we're proxying for, we store the owned names in our own strv
so we can check against them.
* Walk the list of names to check which name would allow the message to
pass, and explicitly use that name as destination of the message. If the
destination is on kdbus, store both the connection's unique name and the
chosen well-known-name in the message. That way, the kernel will make sure
the supplied name is owned by the supplied unique name, at the time of
sending, and return -EREMCHG otherwise.
* Make the policy checks optional by retrieving the bus owner creds, and
when the uid matches the current user's uid and is non-null, don't check
the bus policy.
We need to figure out which of the possible names satisfied the policy,
so we cannot do the iteration in check_policy_item() but have to leave it
to the users.
Test cases amended accordingly.
kdbus learned to accept both a numerical destination ID as well as a
well-known-name. In that case, kdbus makes sure that the numerical ID is in
fact the owner of the provided name and fails otherwise.
This allows for race-free assertion of a bus name owner while sending a
message, which is a requirement for bus-proxyd.
Add two new fields to sd_bus_message, and set the numerical ID to
verify_destination_id if bus_message_setup_kmsg() is called for a
message with a well-known name.
Also, set the destination's name in the kdbus item to .destination_ptr
if it is non-NULL.
Normal users should not touch these fields, and they're not publicy
accessible.
The metadata logic in kdbus has seen a rework, and the only mandatory
change we have to follow for now is that attach_flags in kdbus_cmd_hello
is now split into two parts, attach_flags_send and attach_flags_recv.
Catch up with some changes in kdbus.h:
* KDBUS_{ITEM,ATTACH}_CONN_NAME were renamed to
KDBUS_{ITEM,ATTACH}_CONN_DESCRIPTION, so the term 'name' is not
overloaded as much.
* The item types were re-ordered a little so they are lined up to the
order of the corresponding KDBUS_ATTACH flags
* A new item type KDBUS_ITEM_OWNED_NAME was introduced, designated to
store a struct kdbus_name in item->name. KDBUS_ITEM_NAME soley
stores data in item->str now
* Some kerneldoc fixes
kdbus learned a new ioctl to tell userspace about a bus creator's
credentials, which is what we need to implement sd_bus_get_owner_creds() for
kdbus.
Move the function from sd-bus.c to bus-control.c to be able to reuse
the bus_populate_creds_from_items() helper.
Clean up the function namespace by renaming the following:
sd_bus_get_owner_uid() → sd_bus_get_name_creds_uid()
sd_bus_get_owner_machine_id() → sd_bus_get_name_machine_id()
sd_bus_get_peer_creds() → sd_bus_get_owner_creds()
kdbus learned a new command to query a bus creator's credentials. Sync
kdbus.h first, which also renames some struct to more generic terms.
That is, however, not an ABI break this time.
kdbus_msg.timeout_ns now takes an absolute value, based on CLOCK_MONOTONIC,
in order to eventually support automatically restarted syscalls.
Signed-off-by: Daniel Mack <daniel@zonque.org>
In kdbus.h, the following details changed:
* All commands gained a 'kernel_flags' field to report the flags supported
by the driver. Before, this was done in the 'flags' field in a
bidirectional way, which turned out to be a problem for the code in
sd-bus, as many parts of it reuse the same ioctl struct more than once
and consider them to be owned by userspace.
* Name listings are now returned by a new struct instead of reusing struct
kdbus_cmd_name for that matter. This way, we don't add more unneeded
fields to it and make the API cleaner.
* 'conn_flags' was renamed to 'flags' in struct kdbus_cmd_hello to make
the API a bit more unified.
kdbus.h now has KDBUS_ATTACH_COMM split into KDBUS_ATTACH_TID_COMM and
KDBUS_ATTACH_PID_COMM. The items were split already, so the change in
systemd is easy.
In kdbus.h, the 'features' field has been dropped again. Instead of
negotiating features that way, we decided to make the kernel return the
set of supported flags in each ioctl struct's .flags field, in both the
success and error cases.
Commit 864e17068 ("nspawn: actually allow access to /dev/net/tun in the
container") added "/dev/net/tun" to the list of allowed devices but forgot
to tweak the array length, which caused "/dev/kdbus/*" to be missed.
We need to figure out which of the possible names satisfied the policy,
so we cannot do the iteration in check_policy_item() but have to leave it
to the users.
Test cases amended accordingly.
Sync kdbus.h once again. Two thing have changed:
a) KDBUS_CMD_EP_* was renamed to KDBUS_CMD_ENDPOINT_*
b) struct kdbus_cmd_hello and struct kdbus_cmd_make gained a
'features' bitfield (which is currently unused)
Daniel Mack