picnoir/Systemd
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
Systemd/src/test/test-unit-file.c

931 lines
36 KiB
C
Raw Normal View History

Add SPDX license identifiers to source files under the LGPL This follows what the kernel is doing, c.f. https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=5fd54ace4721fc5ce2bb5aef6318fcf17f421460.
2017-11-18 17:09:20 +01:00
/* SPDX-License-Identifier: LGPL-2.1+ */
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
util-lib: split our string related calls from util.[ch] into its own file string-util.[ch] There are more than enough calls doing string manipulations to deserve its own files, hence do something about it. This patch also sorts the #include blocks of all files that needed to be updated, according to the sorting suggestions from CODING_STYLE. Since pretty much every file needs our string manipulation functions this effectively means that most files have sorted #include blocks now. Also touches a few unrelated include files.
2015-10-24 22:58:24 +02:00
#include <fcntl.h>
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
#include <stddef.h>
util-lib: split our string related calls from util.[ch] into its own file string-util.[ch] There are more than enough calls doing string manipulations to deserve its own files, hence do something about it. This patch also sorts the #include blocks of all files that needed to be updated, according to the sorting suggestions from CODING_STYLE. Since pretty much every file needs our string manipulation functions this effectively means that most files have sorted #include blocks now. Also touches a few unrelated include files.
2015-10-24 22:58:24 +02:00
#include <stdio.h>
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
#include <string.h>
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
#include <sys/capability.h>
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
#include <unistd.h>
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
core: undo the dependency inversion between unit.h and all unit types
2018-05-15 20:17:34 +02:00
#include "all-units.h"
util-lib: split out all temporary file related calls into tmpfiles-util.c This splits out a bunch of functions from fileio.c that have to do with temporary files. Simply to make the header files a bit shorter, and to group things more nicely. No code changes, just some rearranging of source files.
2018-11-30 21:05:27 +01:00
#include "alloc-util.h"
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
#include "capability-util.h"
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
#include "conf-parser.h"
util-lib: split out env file parsing code into env-file.c It's quite complex, let's split this out. No code changes, just some file rearranging.
2018-11-30 22:08:41 +01:00
#include "env-file.h"
util-lib: split out fd-related operations into fd-util.[ch] There are more than enough to deserve their own .c file, hence move them over.
2015-10-25 13:14:12 +01:00
#include "fd-util.h"
fs-util,test: add helper to remove tempfiles This simplifies the use of tempfiles in tests and fixes "leaked" temporary files in test-fileio, test-catalog, test-conf-parser. Not the whole tree is converted.
2018-05-16 11:35:41 +02:00
#include "fs-util.h"
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
#include "hashmap.h"
util-lib: split our string related calls from util.[ch] into its own file string-util.[ch] There are more than enough calls doing string manipulations to deserve its own files, hence do something about it. This patch also sorts the #include blocks of all files that needed to be updated, according to the sorting suggestions from CODING_STYLE. Since pretty much every file needs our string manipulation functions this effectively means that most files have sorted #include blocks now. Also touches a few unrelated include files.
2015-10-24 22:58:24 +02:00
#include "hostname-util.h"
#include "install-printf.h"
#include "install.h"
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
#include "load-fragment.h"
util-lib: split our string related calls from util.[ch] into its own file string-util.[ch] There are more than enough calls doing string manipulations to deserve its own files, hence do something about it. This patch also sorts the #include blocks of all files that needed to be updated, according to the sorting suggestions from CODING_STYLE. Since pretty much every file needs our string manipulation functions this effectively means that most files have sorted #include blocks now. Also touches a few unrelated include files.
2015-10-24 22:58:24 +02:00
#include "macro.h"
tests: override XDG_RUNTIME_DIR where we use the user runtime dir We don#t really support systems where XDG_RUNTIME_DIR is not supported for systemd --user. Hence, let's always set our own XDG_RUNTIME_DIR for tests that involve systemd --user, so that we know it is set, and that it doesn't polute the user's actual runtime dir.
2016-04-08 18:54:05 +02:00
#include "rm-rf.h"
util-lib: split our string related calls from util.[ch] into its own file string-util.[ch] There are more than enough calls doing string manipulations to deserve its own files, hence do something about it. This patch also sorts the #include blocks of all files that needed to be updated, according to the sorting suggestions from CODING_STYLE. Since pretty much every file needs our string manipulation functions this effectively means that most files have sorted #include blocks now. Also touches a few unrelated include files.
2015-10-24 22:58:24 +02:00
#include "specifier.h"
#include "string-util.h"
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
#include "strv.h"
test: Make testing work on systems without or old systemd * Introduce a macro to conditionally execute tests. This avoids skipping the entire test if some parts require systemd * Skip the journal tests when no /etc/machine-id is present * Change test-catalog to load the catalog from the source directory of systemd. * /proc/PID/comm got introduced in v2.6.33 but travis is still using v2.6.32. * Enable make check and make distcheck on the travis build * Use -D"CATALOG_DIR=STR($(abs_top_srcdir)/catalog)" as a STRINGIY would result in the path '/home/ich/source/linux' to be expanded to '/home/ich/source/1' as linux is defined to 1.
2013-07-18 08:30:06 +02:00
#include "test-helper.h"
tests: override XDG_RUNTIME_DIR where we use the user runtime dir We don#t really support systems where XDG_RUNTIME_DIR is not supported for systemd --user. Hence, let's always set our own XDG_RUNTIME_DIR for tests that involve systemd --user, so that we know it is set, and that it doesn't polute the user's actual runtime dir.
2016-04-08 18:54:05 +02:00
#include "tests.h"
util-lib: split out all temporary file related calls into tmpfiles-util.c This splits out a bunch of functions from fileio.c that have to do with temporary files. Simply to make the header files a bit shorter, and to group things more nicely. No code changes, just some rearranging of source files.
2018-11-30 21:05:27 +01:00
#include "tmpfile-util.h"
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
#include "user-util.h"
util-lib: split our string related calls from util.[ch] into its own file string-util.[ch] There are more than enough calls doing string manipulations to deserve its own files, hence do something about it. This patch also sorts the #include blocks of all files that needed to be updated, according to the sorting suggestions from CODING_STYLE. Since pretty much every file needs our string manipulation functions this effectively means that most files have sorted #include blocks now. Also touches a few unrelated include files.
2015-10-24 22:58:24 +02:00
#include "util.h"
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
test-unit-file: return error without dumping core on permission error
2013-07-30 15:07:20 +02:00
static int test_unit_file_get_set(void) {
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
int r;
Hashmap *h;
Iterator i;
UnitFileList *p;
hashmap: introduce hash_ops to make struct Hashmap smaller It is redundant to store 'hash' and 'compare' function pointers in struct Hashmap separately. The functions always comprise a pair. Store a single pointer to struct hash_ops instead. systemd keeps hundreds of hashmaps, so this saves a little bit of memory.
2014-08-13 01:00:18 +02:00
h = hashmap_new(&string_hash_ops);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(h);
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
core: Filter by unit name behind the D-Bus, instead on the client side (#3142) This commit improves systemd performance on the systems which have thousands of units.
2016-04-29 15:59:51 +02:00
r = unit_file_get_list(UNIT_FILE_SYSTEM, NULL, h, NULL, NULL);
tests: add a helper function to skip with errno
2018-09-13 16:11:16 +02:00
if (IN_SET(r, -EPERM, -EACCES))
return log_tests_skipped_errno(r, "unit_file_get_list");
test-unit-file: skip if unit_file_get_list returns permission denied
2014-05-24 18:16:42 +02:00
tests: get rid of strerror
2016-09-12 20:55:34 +02:00
log_full_errno(r == 0 ? LOG_INFO : LOG_ERR, r,
"unit_file_get_list: %m");
test-unit-file: return error without dumping core on permission error
2013-07-30 15:07:20 +02:00
if (r < 0)
return EXIT_FAILURE;
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
HASHMAP_FOREACH(p, h, i)
printf("%s = %s\n", p->path, unit_file_state_to_string(p->state));
unit_file_list_free(h);
test-unit-file: return error without dumping core on permission error
2013-07-30 15:07:20 +02:00
return 0;
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
}
static void check_execcommand(ExecCommand *c,
const char* path,
const char* argv0,
const char* argv1,
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
const char* argv2,
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
bool ignore) {
test-unit-file: don't access out-of-bounds memory Fixes an error introduced by me when the test was added.
2015-01-11 22:52:50 +01:00
size_t n;
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(c);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("expect: \"%s\" [\"%s\" \"%s\" \"%s\"]",
path, argv0 ?: path, argv1, argv2);
test-unit-file: don't access out-of-bounds memory Fixes an error introduced by me when the test was added.
2015-01-11 22:52:50 +01:00
n = strv_length(c->argv);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("actual: \"%s\" [\"%s\" \"%s\" \"%s\"]",
test-unit-file: don't access out-of-bounds memory Fixes an error introduced by me when the test was added.
2015-01-11 22:52:50 +01:00
c->path, c->argv[0], n > 0 ? c->argv[1] : NULL, n > 1 ? c->argv[2] : NULL);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(streq(c->path, path));
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(streq(c->argv[0], argv0 ?: path));
test-unit-file: don't access out-of-bounds memory Fixes an error introduced by me when the test was added.
2015-01-11 22:52:50 +01:00
if (n > 0)
assert_se(streq_ptr(c->argv[1], argv1));
if (n > 1)
assert_se(streq_ptr(c->argv[2], argv2));
execute: replace command flag bools by a flags field This way, we can extend it later on in an easier way, and can pass it along nicely.
2017-08-01 10:16:42 +02:00
assert_se(!!(c->flags & EXEC_COMMAND_IGNORE_FAILURE) == ignore);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
}
static void test_config_parse_exec(void) {
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
/* int config_parse_exec(
util: Fix assertion in split() on missing ' When parsing a unit with a trailing slash after an escaped line break, like ExecStart=/bin/echo 'foo \ bar' the split() function (through config_parse()) asserted and crashed pid 1: Assertion 'current[*l + 1] == quotechars[0]' failed at ../src/shared/util.c:583, function split(). Aborting. Fix this by returning an error in this case ("trailing garbage"). Add corresponding test case. Also fix the missing "unit" argument of config_parse_exec() in the comment. https://launchpad.net/bugs/1447243
2015-04-23 00:09:43 +02:00
const char *unit,
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
const char *filename,
unsigned line,
const char *section,
unsigned section_line,
const char *lvalue,
int ltype,
const char *rvalue,
void *data,
void *userdata) */
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
int r;
ExecCommand *c = NULL, *c1;
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
const char *ccc;
Introduce _cleanup_(manager_freep)
2018-03-09 21:55:55 +01:00
_cleanup_(manager_freep) Manager *m = NULL;
Introduce _cleanup_(unit_freep)
2018-03-09 21:34:28 +01:00
_cleanup_(unit_freep) Unit *u = NULL;
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
Make test_run into a flags field and disable generators again Now generators are only run in systemd --test mode, where this makes most sense (how are you going to test what would happen otherwise?). Fixes #6842. v2: - rename test_run to test_run_flags
2017-09-16 11:19:43 +02:00
r = manager_new(UNIT_FILE_USER, MANAGER_TEST_RUN_MINIMAL, &m);
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
if (MANAGER_SKIP_TEST(r)) {
tests: get rid of strerror
2016-09-12 20:55:34 +02:00
log_notice_errno(r, "Skipping test: manager_new: %m");
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
return;
}
assert_se(r >= 0);
assert_se(manager_startup(m, NULL, NULL) >= 0);
assert_se(u = unit_new(m, sizeof(Service)));
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* basic test */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 1, "section", 1,
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
"LValue", 0, "/RValue r1",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(r >= 0);
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
check_execcommand(c, "/RValue", "/RValue", "r1", NULL, false);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 2, "section", 1,
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
"LValue", 0, "/RValue///slashes r1///",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* test slashes */");
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(r >= 0);
c1 = c->command_next;
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
check_execcommand(c1, "/RValue/slashes", "/RValue///slashes", "r1///", NULL, false);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* trailing slash */");
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, "/RValue/ argv0 r1",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core/load-fragment: refuse units with errors in certain directives If an error is encountered in any of the Exec* lines, WorkingDirectory, SELinuxContext, ApparmorProfile, SmackProcessLabel, Service (in .socket units), User, or Group, refuse to load the unit. If the config stanza has support, ignore the failure if '-' is present. For those configuration directives, even if we started the unit, it's pretty likely that it'll do something unexpected (like write files in a wrong place, or with a wrong context, or run with wrong permissions, etc). It seems better to refuse to start the unit and have the admin clean up the configuration without giving the service a chance to mess up stuff. Note that all "security" options that restrict what the unit can do (Capabilities, AmbientCapabilities, Restrict*, SystemCallFilter, Limit*, PrivateDevices, Protect*, etc) are _not_ treated like this. Such options are only supplementary, and are not always available depending on the architecture and compilation options, so unit authors have to make sure that the service runs correctly without them anyway. Fixes #6237, #6277.
2017-07-06 19:28:19 +02:00
assert_se(r == -ENOEXEC);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(c1->command_next == NULL);
log_info("/* honour_argv0 */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 3, "section", 1,
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
"LValue", 0, "@/RValue///slashes2 ///argv0 r1",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
check_execcommand(c1, "/RValue/slashes2", "///argv0", "r1", NULL, false);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
core/load-fragment: avoid allocating 0 bytes when given an invalid command With a command line like "@/something" we would allocate an array with 0 elements. Avoid that, and add a test too.
2015-01-11 22:57:02 +01:00
log_info("/* honour_argv0, no args */");
r = config_parse_exec(NULL, "fake", 3, "section", 1,
"LValue", 0, "@/RValue",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core/load-fragment: refuse units with errors in certain directives If an error is encountered in any of the Exec* lines, WorkingDirectory, SELinuxContext, ApparmorProfile, SmackProcessLabel, Service (in .socket units), User, or Group, refuse to load the unit. If the config stanza has support, ignore the failure if '-' is present. For those configuration directives, even if we started the unit, it's pretty likely that it'll do something unexpected (like write files in a wrong place, or with a wrong context, or run with wrong permissions, etc). It seems better to refuse to start the unit and have the admin clean up the configuration without giving the service a chance to mess up stuff. Note that all "security" options that restrict what the unit can do (Capabilities, AmbientCapabilities, Restrict*, SystemCallFilter, Limit*, PrivateDevices, Protect*, etc) are _not_ treated like this. Such options are only supplementary, and are not always available depending on the architecture and compilation options, so unit authors have to make sure that the service runs correctly without them anyway. Fixes #6237, #6277.
2017-07-06 19:28:19 +02:00
assert_se(r == -ENOEXEC);
core/load-fragment: avoid allocating 0 bytes when given an invalid command With a command line like "@/something" we would allocate an array with 0 elements. Avoid that, and add a test too.
2015-01-11 22:57:02 +01:00
assert_se(c1->command_next == NULL);
load-fragment: reset the list on an ExecStart= containing only whitespace This is consistent with how an empty string works in an ExecStart= statement. We should not differentiate between an empty string and whitespace only (since they look the same.) Update the test case with whitespace only to reflect that the list is reset. Tested that `test-unit-file` passes and other test cases are not affected. Installed the patched systemd binaries on a machine, booted it, looked for out of the usual behavior but did not find any.
2015-06-06 08:12:25 +02:00
log_info("/* no command, whitespace only, reset */");
core/load-fragment: avoid allocating 0 bytes when given an invalid command With a command line like "@/something" we would allocate an array with 0 elements. Avoid that, and add a test too.
2015-01-11 22:57:02 +01:00
r = config_parse_exec(NULL, "fake", 3, "section", 1,
"LValue", 0, " ",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core/load-fragment: avoid allocating 0 bytes when given an invalid command With a command line like "@/something" we would allocate an array with 0 elements. Avoid that, and add a test too.
2015-01-11 22:57:02 +01:00
assert_se(r == 0);
load-fragment: reset the list on an ExecStart= containing only whitespace This is consistent with how an empty string works in an ExecStart= statement. We should not differentiate between an empty string and whitespace only (since they look the same.) Update the test case with whitespace only to reflect that the list is reset. Tested that `test-unit-file` passes and other test cases are not affected. Installed the patched systemd binaries on a machine, booted it, looked for out of the usual behavior but did not find any.
2015-06-06 08:12:25 +02:00
assert_se(c == NULL);
core/load-fragment: avoid allocating 0 bytes when given an invalid command With a command line like "@/something" we would allocate an array with 0 elements. Avoid that, and add a test too.
2015-01-11 22:57:02 +01:00
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* ignore && honour_argv0 */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 4, "section", 1,
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
"LValue", 0, "-@/RValue///slashes3 argv0a r1",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(r >= 0);
load-fragment: reset the list on an ExecStart= containing only whitespace This is consistent with how an empty string works in an ExecStart= statement. We should not differentiate between an empty string and whitespace only (since they look the same.) Update the test case with whitespace only to reflect that the list is reset. Tested that `test-unit-file` passes and other test cases are not affected. Installed the patched systemd binaries on a machine, booted it, looked for out of the usual behavior but did not find any.
2015-06-06 08:12:25 +02:00
c1 = c;
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
check_execcommand(c1, "/RValue/slashes3", "argv0a", "r1", NULL, true);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* ignore && honour_argv0 */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 4, "section", 1,
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
"LValue", 0, "@-/RValue///slashes4 argv0b r1",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core: lift restriction on order of - and @ in ExecStart
2012-11-15 15:25:05 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
check_execcommand(c1, "/RValue/slashes4", "argv0b", "r1", NULL, true);
core: lift restriction on order of - and @ in ExecStart
2012-11-15 15:25:05 +01:00
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* ignore && ignore */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 4, "section", 1,
core: lift restriction on order of - and @ in ExecStart
2012-11-15 15:25:05 +01:00
"LValue", 0, "--/RValue argv0 r1",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core: lift restriction on order of - and @ in ExecStart
2012-11-15 15:25:05 +01:00
assert_se(r == 0);
assert_se(c1->command_next == NULL);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* ignore && ignore (2) */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 4, "section", 1,
core: lift restriction on order of - and @ in ExecStart
2012-11-15 15:25:05 +01:00
"LValue", 0, "-@-/RValue argv0 r1",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core: lift restriction on order of - and @ in ExecStart
2012-11-15 15:25:05 +01:00
assert_se(r == 0);
assert_se(c1->command_next == NULL);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* semicolon */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 5, "section", 1,
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
"LValue", 0,
"-@/RValue argv0 r1 ; "
"/goo/goo boo",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
check_execcommand(c1, "/RValue", "argv0", "r1", NULL, true);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
c1 = c1->command_next;
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
check_execcommand(c1, "/goo/goo", NULL, "boo", NULL, false);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
log_info("/* two semicolons in a row */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"-@/RValue argv0 r1 ; ; "
"/goo/goo boo",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core/load-fragment: refuse units with errors in certain directives If an error is encountered in any of the Exec* lines, WorkingDirectory, SELinuxContext, ApparmorProfile, SmackProcessLabel, Service (in .socket units), User, or Group, refuse to load the unit. If the config stanza has support, ignore the failure if '-' is present. For those configuration directives, even if we started the unit, it's pretty likely that it'll do something unexpected (like write files in a wrong place, or with a wrong context, or run with wrong permissions, etc). It seems better to refuse to start the unit and have the admin clean up the configuration without giving the service a chance to mess up stuff. Note that all "security" options that restrict what the unit can do (Capabilities, AmbientCapabilities, Restrict*, SystemCallFilter, Limit*, PrivateDevices, Protect*, etc) are _not_ treated like this. Such options are only supplementary, and are not always available depending on the architecture and compilation options, so unit authors have to make sure that the service runs correctly without them anyway. Fixes #6237, #6277.
2017-07-06 19:28:19 +02:00
assert_se(r == -ENOEXEC);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
c1 = c1->command_next;
check_execcommand(c1, "/RValue", "argv0", "r1", NULL, true);
/* second command fails because the executable name is ";" */
assert_se(c1->command_next == NULL);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* trailing semicolon */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 5, "section", 1,
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
"LValue", 0,
"-@/RValue argv0 r1 ; ",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
check_execcommand(c1, "/RValue", "argv0", "r1", NULL, true);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
assert_se(c1->command_next == NULL);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
log_info("/* trailing semicolon, no whitespace */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"-@/RValue argv0 r1 ;",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1, "/RValue", "argv0", "r1", NULL, true);
assert_se(c1->command_next == NULL);
log_info("/* trailing semicolon in single quotes */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"-@/RValue argv0 r1 ';'",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
assert_se(r >= 0);
c1 = c1->command_next;
load-fragment: use unquote_first_word in config_parse_exec Convert config_parse_exec() from using FOREACH_WORD_QUOTED into a loop of unquote_first_word. Loop through the arguments only once (the FOREACH_WORD_QUOTED implementation did it twice, once to count them and another time to process and store them.) Use newly introduced flag UNQUOTE_UNESCAPE_RELAX to preserve unrecognized escape sequences such as regexps matches such as "\w", "\d", etc. (Valid escape sequences such as "\s" or "\b" still need an extra backslash if literals are desired for regexps.) Differences in behavior: - Handle ; (command separator) in special, so that only ; on its own is valid for that purpose, an quoted semicolon ";" or ';' will now behave as a literal semicolon. This is probably what was initially intended. - Handle \; (to introduce a literal semicolon) in special, so that only \; is turned into a semicolon but not \\; or "\\;" or "\;" which are kept as a literal \; in the output. This is probably what was initially intended. Known issues: - Using an empty string (for example, ExecStartPre=<empty>) will empty the list and remove the existing commands, but using whitespace only (for example, ExecStartPre=<spaces>) will not. This is a pre-existing issue and will be dealt with in a follow up commit. Tested: - Unit tests passing. Also `make distcheck` still works as expected. - Installed it on a local machine and booted with it, checked console output, systemctl and journalctl output, did not notice any issues running the patched systemd binaries. Relevant bug: https://bugs.freedesktop.org/show_bug.cgi?id=90794
2015-05-31 07:48:52 +02:00
check_execcommand(c1, "/RValue", "argv0", "r1", ";", true);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* escaped semicolon */");
conf-parser: distinguish between multiple sections with the same name Pass on the line on which a section was decleared to the parsers, so they can distinguish between multiple sections (if they chose to). Currently no parsers take advantage of this, but a follow-up patch will do that to distinguish [Address] Address=192.168.0.1/24 Label=one [Address] Address=192.168.0.2/24 Label=two from [Address] Address=192.168.0.1/24 Label=one Address=192.168.0.2/24 Label=two
2013-11-19 16:17:55 +01:00
r = config_parse_exec(NULL, "fake", 5, "section", 1,
core: interpret \; token in ExecStart as escaped ; Some commands (like 'find') take a semicolon as separate arg. With current parser implementation there is no way to pass one. Patch adds token \;
2012-11-03 20:52:02 +01:00
"LValue", 0,
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
"/bin/find \\;",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
check_execcommand(c1, "/bin/find", NULL, ";", NULL, false);
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* escaped semicolon with following arg */");
test-unit-file: add test for semicolon escaping https://bugs.freedesktop.org/show_bug.cgi?id=87393
2014-12-18 06:52:28 +01:00
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
"/sbin/find \\; /x",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1,
"/sbin/find", NULL, ";", "/x", false);
log_info("/* escaped semicolon as part of an expression */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"/sbin/find \\;x",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core: interpret \; token in ExecStart as escaped ; Some commands (like 'find') take a semicolon as separate arg. With current parser implementation there is no way to pass one. Patch adds token \;
2012-11-03 20:52:02 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1,
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
"/sbin/find", NULL, "\\;x", NULL, false);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
test-unit-file: add test for improperly escaped exec string Add a regression test for the recent breakage of handling improperly escaped exec strings in unit files. Code contributed by Martin Pitt: https://bugs.freedesktop.org/show_bug.cgi?id=90794
2015-06-01 18:05:02 +02:00
log_info("/* encoded semicolon */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"/bin/find \\073",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
test-unit-file: add test for improperly escaped exec string Add a regression test for the recent breakage of handling improperly escaped exec strings in unit files. Code contributed by Martin Pitt: https://bugs.freedesktop.org/show_bug.cgi?id=90794
2015-06-01 18:05:02 +02:00
assert_se(r >= 0);
c1 = c1->command_next;
test-unit-file.c: fixup the test for commit 3b51f8ddd5
2015-06-03 16:36:20 +02:00
check_execcommand(c1, "/bin/find", NULL, ";", NULL, false);
test-unit-file: add test for improperly escaped exec string Add a regression test for the recent breakage of handling improperly escaped exec strings in unit files. Code contributed by Martin Pitt: https://bugs.freedesktop.org/show_bug.cgi?id=90794
2015-06-01 18:05:02 +02:00
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
log_info("/* quoted semicolon */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"/bin/find \";\"",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
assert_se(r >= 0);
c1 = c1->command_next;
load-fragment: use unquote_first_word in config_parse_exec Convert config_parse_exec() from using FOREACH_WORD_QUOTED into a loop of unquote_first_word. Loop through the arguments only once (the FOREACH_WORD_QUOTED implementation did it twice, once to count them and another time to process and store them.) Use newly introduced flag UNQUOTE_UNESCAPE_RELAX to preserve unrecognized escape sequences such as regexps matches such as "\w", "\d", etc. (Valid escape sequences such as "\s" or "\b" still need an extra backslash if literals are desired for regexps.) Differences in behavior: - Handle ; (command separator) in special, so that only ; on its own is valid for that purpose, an quoted semicolon ";" or ';' will now behave as a literal semicolon. This is probably what was initially intended. - Handle \; (to introduce a literal semicolon) in special, so that only \; is turned into a semicolon but not \\; or "\\;" or "\;" which are kept as a literal \; in the output. This is probably what was initially intended. Known issues: - Using an empty string (for example, ExecStartPre=<empty>) will empty the list and remove the existing commands, but using whitespace only (for example, ExecStartPre=<spaces>) will not. This is a pre-existing issue and will be dealt with in a follow up commit. Tested: - Unit tests passing. Also `make distcheck` still works as expected. - Installed it on a local machine and booted with it, checked console output, systemctl and journalctl output, did not notice any issues running the patched systemd binaries. Relevant bug: https://bugs.freedesktop.org/show_bug.cgi?id=90794
2015-05-31 07:48:52 +02:00
check_execcommand(c1, "/bin/find", NULL, ";", NULL, false);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
log_info("/* quoted semicolon with following arg */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"/sbin/find \";\" /x",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1,
load-fragment: use unquote_first_word in config_parse_exec Convert config_parse_exec() from using FOREACH_WORD_QUOTED into a loop of unquote_first_word. Loop through the arguments only once (the FOREACH_WORD_QUOTED implementation did it twice, once to count them and another time to process and store them.) Use newly introduced flag UNQUOTE_UNESCAPE_RELAX to preserve unrecognized escape sequences such as regexps matches such as "\w", "\d", etc. (Valid escape sequences such as "\s" or "\b" still need an extra backslash if literals are desired for regexps.) Differences in behavior: - Handle ; (command separator) in special, so that only ; on its own is valid for that purpose, an quoted semicolon ";" or ';' will now behave as a literal semicolon. This is probably what was initially intended. - Handle \; (to introduce a literal semicolon) in special, so that only \; is turned into a semicolon but not \\; or "\\;" or "\;" which are kept as a literal \; in the output. This is probably what was initially intended. Known issues: - Using an empty string (for example, ExecStartPre=<empty>) will empty the list and remove the existing commands, but using whitespace only (for example, ExecStartPre=<spaces>) will not. This is a pre-existing issue and will be dealt with in a follow up commit. Tested: - Unit tests passing. Also `make distcheck` still works as expected. - Installed it on a local machine and booted with it, checked console output, systemctl and journalctl output, did not notice any issues running the patched systemd binaries. Relevant bug: https://bugs.freedesktop.org/show_bug.cgi?id=90794
2015-05-31 07:48:52 +02:00
"/sbin/find", NULL, ";", "/x", false);
tests: additional cases in test-unit-file These tests will be useful to check the cases regarding quoted and escaped semicolon when we switch to using unquote_first_word. Additionally, convert some of the tests that have semicolons so that the argument after the semicolon looks like a path (starts with /) so that we can see the change of behavior when making config_parse_exec more strict about what it accepts as a command separator.
2015-06-09 07:01:52 +02:00
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* spaces in the filename */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"\"/PATH WITH SPACES/daemon\" -1 -2",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1,
"/PATH WITH SPACES/daemon", NULL, "-1", "-2", false);
log_info("/* spaces in the filename, no args */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"\"/PATH WITH SPACES/daemon -1 -2\"",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1,
"/PATH WITH SPACES/daemon -1 -2", NULL, NULL, NULL, false);
log_info("/* spaces in the filename, everything quoted */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"\"/PATH WITH SPACES/daemon\" \"-1\" '-2'",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1,
"/PATH WITH SPACES/daemon", NULL, "-1", "-2", false);
log_info("/* escaped spaces in the filename */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"\"/PATH\\sWITH\\sSPACES/daemon\" '-1 -2'",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1,
"/PATH WITH SPACES/daemon", NULL, "-1 -2", NULL, false);
log_info("/* escaped spaces in the filename (2) */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"\"/PATH\\x20WITH\\x20SPACES/daemon\" \"-1 -2\"",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1,
"/PATH WITH SPACES/daemon", NULL, "-1 -2", NULL, false);
for (ccc = "abfnrtv\\\'\"x"; *ccc; ccc++) {
/* \\x is an incomplete hexadecimal sequence, invalid because of the slash */
char path[] = "/path\\X";
path[sizeof(path) - 2] = *ccc;
log_info("/* invalid character: \\%c */", *ccc);
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, path,
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core/load-fragment: refuse units with errors in certain directives If an error is encountered in any of the Exec* lines, WorkingDirectory, SELinuxContext, ApparmorProfile, SmackProcessLabel, Service (in .socket units), User, or Group, refuse to load the unit. If the config stanza has support, ignore the failure if '-' is present. For those configuration directives, even if we started the unit, it's pretty likely that it'll do something unexpected (like write files in a wrong place, or with a wrong context, or run with wrong permissions, etc). It seems better to refuse to start the unit and have the admin clean up the configuration without giving the service a chance to mess up stuff. Note that all "security" options that restrict what the unit can do (Capabilities, AmbientCapabilities, Restrict*, SystemCallFilter, Limit*, PrivateDevices, Protect*, etc) are _not_ treated like this. Such options are only supplementary, and are not always available depending on the architecture and compilation options, so unit authors have to make sure that the service runs correctly without them anyway. Fixes #6237, #6277.
2017-07-06 19:28:19 +02:00
assert_se(r == -ENOEXEC);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(c1->command_next == NULL);
}
log_info("/* valid character: \\s */");
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, "/path\\s",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1, "/path ", NULL, NULL, NULL, false);
test-unit-file: add test for improperly escaped exec string Add a regression test for the recent breakage of handling improperly escaped exec strings in unit files. Code contributed by Martin Pitt: https://bugs.freedesktop.org/show_bug.cgi?id=90794
2015-06-01 18:05:02 +02:00
log_info("/* quoted backslashes */");
r = config_parse_exec(NULL, "fake", 5, "section", 1,
"LValue", 0,
"/bin/grep '\\w+\\K'",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
test-unit-file: add test for improperly escaped exec string Add a regression test for the recent breakage of handling improperly escaped exec strings in unit files. Code contributed by Martin Pitt: https://bugs.freedesktop.org/show_bug.cgi?id=90794
2015-06-01 18:05:02 +02:00
assert_se(r >= 0);
c1 = c1->command_next;
check_execcommand(c1, "/bin/grep", NULL, "\\w+\\K", NULL, false);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
log_info("/* trailing backslash: \\ */");
/* backslash is invalid */
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, "/path\\",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core/load-fragment: refuse units with errors in certain directives If an error is encountered in any of the Exec* lines, WorkingDirectory, SELinuxContext, ApparmorProfile, SmackProcessLabel, Service (in .socket units), User, or Group, refuse to load the unit. If the config stanza has support, ignore the failure if '-' is present. For those configuration directives, even if we started the unit, it's pretty likely that it'll do something unexpected (like write files in a wrong place, or with a wrong context, or run with wrong permissions, etc). It seems better to refuse to start the unit and have the admin clean up the configuration without giving the service a chance to mess up stuff. Note that all "security" options that restrict what the unit can do (Capabilities, AmbientCapabilities, Restrict*, SystemCallFilter, Limit*, PrivateDevices, Protect*, etc) are _not_ treated like this. Such options are only supplementary, and are not always available depending on the architecture and compilation options, so unit authors have to make sure that the service runs correctly without them anyway. Fixes #6237, #6277.
2017-07-06 19:28:19 +02:00
assert_se(r == -ENOEXEC);
load-fragment: allow quoting in command name and document allowed escapes The handling of the command name and other arguments is unified. This simplifies things and should make them more predictable for users. Incidentally, this makes ExecStart handling match the .desktop file specification, apart for the requirment for an absolute path. https://bugs.freedesktop.org/show_bug.cgi?id=86171
2014-12-19 00:08:13 +01:00
assert_se(c1->command_next == NULL);
core: interpret \; token in ExecStart as escaped ; Some commands (like 'find') take a semicolon as separate arg. With current parser implementation there is no way to pass one. Patch adds token \;
2012-11-03 20:52:02 +01:00
util: Fix assertion in split() on missing ' When parsing a unit with a trailing slash after an escaped line break, like ExecStart=/bin/echo 'foo \ bar' the split() function (through config_parse()) asserted and crashed pid 1: Assertion 'current[*l + 1] == quotechars[0]' failed at ../src/shared/util.c:583, function split(). Aborting. Fix this by returning an error in this case ("trailing garbage"). Add corresponding test case. Also fix the missing "unit" argument of config_parse_exec() in the comment. https://launchpad.net/bugs/1447243
2015-04-23 00:09:43 +02:00
log_info("/* missing ending ' */");
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, "/path 'foo",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core/load-fragment: refuse units with errors in certain directives If an error is encountered in any of the Exec* lines, WorkingDirectory, SELinuxContext, ApparmorProfile, SmackProcessLabel, Service (in .socket units), User, or Group, refuse to load the unit. If the config stanza has support, ignore the failure if '-' is present. For those configuration directives, even if we started the unit, it's pretty likely that it'll do something unexpected (like write files in a wrong place, or with a wrong context, or run with wrong permissions, etc). It seems better to refuse to start the unit and have the admin clean up the configuration without giving the service a chance to mess up stuff. Note that all "security" options that restrict what the unit can do (Capabilities, AmbientCapabilities, Restrict*, SystemCallFilter, Limit*, PrivateDevices, Protect*, etc) are _not_ treated like this. Such options are only supplementary, and are not always available depending on the architecture and compilation options, so unit authors have to make sure that the service runs correctly without them anyway. Fixes #6237, #6277.
2017-07-06 19:28:19 +02:00
assert_se(r == -ENOEXEC);
util: Fix assertion in split() on missing ' When parsing a unit with a trailing slash after an escaped line break, like ExecStart=/bin/echo 'foo \ bar' the split() function (through config_parse()) asserted and crashed pid 1: Assertion 'current[*l + 1] == quotechars[0]' failed at ../src/shared/util.c:583, function split(). Aborting. Fix this by returning an error in this case ("trailing garbage"). Add corresponding test case. Also fix the missing "unit" argument of config_parse_exec() in the comment. https://launchpad.net/bugs/1447243
2015-04-23 00:09:43 +02:00
assert_se(c1->command_next == NULL);
log_info("/* missing ending ' with trailing backslash */");
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, "/path 'foo\\",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core/load-fragment: refuse units with errors in certain directives If an error is encountered in any of the Exec* lines, WorkingDirectory, SELinuxContext, ApparmorProfile, SmackProcessLabel, Service (in .socket units), User, or Group, refuse to load the unit. If the config stanza has support, ignore the failure if '-' is present. For those configuration directives, even if we started the unit, it's pretty likely that it'll do something unexpected (like write files in a wrong place, or with a wrong context, or run with wrong permissions, etc). It seems better to refuse to start the unit and have the admin clean up the configuration without giving the service a chance to mess up stuff. Note that all "security" options that restrict what the unit can do (Capabilities, AmbientCapabilities, Restrict*, SystemCallFilter, Limit*, PrivateDevices, Protect*, etc) are _not_ treated like this. Such options are only supplementary, and are not always available depending on the architecture and compilation options, so unit authors have to make sure that the service runs correctly without them anyway. Fixes #6237, #6277.
2017-07-06 19:28:19 +02:00
assert_se(r == -ENOEXEC);
util: Fix assertion in split() on missing ' When parsing a unit with a trailing slash after an escaped line break, like ExecStart=/bin/echo 'foo \ bar' the split() function (through config_parse()) asserted and crashed pid 1: Assertion 'current[*l + 1] == quotechars[0]' failed at ../src/shared/util.c:583, function split(). Aborting. Fix this by returning an error in this case ("trailing garbage"). Add corresponding test case. Also fix the missing "unit" argument of config_parse_exec() in the comment. https://launchpad.net/bugs/1447243
2015-04-23 00:09:43 +02:00
assert_se(c1->command_next == NULL);
core: Fix assertion with empty Exec*= paths An Exec*= line with whitespace after modifiers, like ExecStart=- /bin/true is considered to have an empty command path. This is as specified, but causes systemd to crash with Assertion 'skip < l' failed at ../src/core/load-fragment.c:607, function config_parse_exec(). Aborting. Aborted (core dumped) Fix this by logging an error instead and ignoring the invalid line. Add corresponding test cases. Also add a test case for a completely empty value which resets the command list. https://launchpad.net/bugs/1454173
2015-05-14 09:06:40 +02:00
log_info("/* invalid space between modifiers */");
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, "- /path",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core: Fix assertion with empty Exec*= paths An Exec*= line with whitespace after modifiers, like ExecStart=- /bin/true is considered to have an empty command path. This is as specified, but causes systemd to crash with Assertion 'skip < l' failed at ../src/core/load-fragment.c:607, function config_parse_exec(). Aborting. Aborted (core dumped) Fix this by logging an error instead and ignoring the invalid line. Add corresponding test cases. Also add a test case for a completely empty value which resets the command list. https://launchpad.net/bugs/1454173
2015-05-14 09:06:40 +02:00
assert_se(r == 0);
assert_se(c1->command_next == NULL);
log_info("/* only modifiers, no path */");
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, "-",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core: Fix assertion with empty Exec*= paths An Exec*= line with whitespace after modifiers, like ExecStart=- /bin/true is considered to have an empty command path. This is as specified, but causes systemd to crash with Assertion 'skip < l' failed at ../src/core/load-fragment.c:607, function config_parse_exec(). Aborting. Aborted (core dumped) Fix this by logging an error instead and ignoring the invalid line. Add corresponding test cases. Also add a test case for a completely empty value which resets the command list. https://launchpad.net/bugs/1454173
2015-05-14 09:06:40 +02:00
assert_se(r == 0);
assert_se(c1->command_next == NULL);
log_info("/* empty argument, reset */");
r = config_parse_exec(NULL, "fake", 4, "section", 1,
"LValue", 0, "",
test: pass temporary allocated unit to config_parse_exec() Now, config_parse_exec() use the userdata as unit and assert on NULL. So allocate empty unit and use that on config_parse_exec().
2016-01-12 08:40:52 +01:00
&c, u);
core: Fix assertion with empty Exec*= paths An Exec*= line with whitespace after modifiers, like ExecStart=- /bin/true is considered to have an empty command path. This is as specified, but causes systemd to crash with Assertion 'skip < l' failed at ../src/core/load-fragment.c:607, function config_parse_exec(). Aborting. Aborted (core dumped) Fix this by logging an error instead and ignoring the invalid line. Add corresponding test cases. Also add a test case for a completely empty value which resets the command list. https://launchpad.net/bugs/1454173
2015-05-14 09:06:40 +02:00
assert_se(r == 0);
assert_se(c == NULL);
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
exec_command_free_list(c);
}
test-unit-file: add test for config_parse_log_extra_fields()
2017-11-14 14:58:37 +01:00
static void test_config_parse_log_extra_fields(void) {
/* int config_parse_log_extra_fields(
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 r;
Introduce _cleanup_(manager_freep)
2018-03-09 21:55:55 +01:00
_cleanup_(manager_freep) Manager *m = NULL;
Introduce _cleanup_(unit_freep)
2018-03-09 21:34:28 +01:00
_cleanup_(unit_freep) Unit *u = NULL;
test-unit-file: add test for config_parse_log_extra_fields()
2017-11-14 14:58:37 +01:00
ExecContext c = {};
r = manager_new(UNIT_FILE_USER, MANAGER_TEST_RUN_MINIMAL, &m);
if (MANAGER_SKIP_TEST(r)) {
log_notice_errno(r, "Skipping test: manager_new: %m");
return;
}
assert_se(r >= 0);
assert_se(manager_startup(m, NULL, NULL) >= 0);
assert_se(u = unit_new(m, sizeof(Service)));
log_info("/* %s – basic test */", __func__);
r = config_parse_log_extra_fields(NULL, "fake", 1, "section", 1,
"LValue", 0, "FOO=BAR \"QOOF=quux ' ' \"",
&c, u);
assert_se(r >= 0);
assert_se(c.n_log_extra_fields == 2);
assert_se(strneq(c.log_extra_fields[0].iov_base, "FOO=BAR", c.log_extra_fields[0].iov_len));
assert_se(strneq(c.log_extra_fields[1].iov_base, "QOOF=quux ' ' ", c.log_extra_fields[1].iov_len));
log_info("/* %s – add some */", __func__);
r = config_parse_log_extra_fields(NULL, "fake", 1, "section", 1,
"LValue", 0, "FOO2=BAR2 QOOF2=quux ' '",
&c, u);
assert_se(r >= 0);
assert_se(c.n_log_extra_fields == 4);
assert_se(strneq(c.log_extra_fields[0].iov_base, "FOO=BAR", c.log_extra_fields[0].iov_len));
assert_se(strneq(c.log_extra_fields[1].iov_base, "QOOF=quux ' ' ", c.log_extra_fields[1].iov_len));
assert_se(strneq(c.log_extra_fields[2].iov_base, "FOO2=BAR2", c.log_extra_fields[2].iov_len));
assert_se(strneq(c.log_extra_fields[3].iov_base, "QOOF2=quux", c.log_extra_fields[3].iov_len));
exec_context_dump(&c, stdout, " --> ");
log_info("/* %s – reset */", __func__);
r = config_parse_log_extra_fields(NULL, "fake", 1, "section", 1,
"LValue", 0, "",
&c, u);
assert_se(r >= 0);
assert_se(c.n_log_extra_fields == 0);
exec_context_free_log_extra_fields(&c);
log_info("/* %s – bye */", __func__);
}
shared: rework env file reader Implement this with a proper state machine, so that newlines and escaped chars can appear in string assignments. This should bring the parser much closer to shell.
2013-04-03 19:04:03 +02:00
#define env_file_1 \
"a=a\n" \
"b=b\\\n" \
"c\n" \
"d=d\\\n" \
"e\\\n" \
"f\n" \
"g=g\\ \n" \
"h=h\n" \
"i=i\\"
#define env_file_2 \
"a=a\\\n"
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
tests: add test for continuation lines with comments
2013-02-13 05:28:48 +01:00
#define env_file_3 \
"#SPAMD_ARGS=\"-d --socketpath=/var/lib/bulwark/spamd \\\n" \
"#--nouser-config \\\n" \
"normal=line"
fileio: in envfiles, do not skip lines following empty lines https://bugs.freedesktop.org/show_bug.cgi?id=63477
2013-04-14 02:22:53 +02:00
#define env_file_4 \
"# Generated\n" \
"\n" \
"HWMON_MODULES=\"coretemp f71882fg\"\n" \
"\n" \
"# For compatibility reasons\n" \
"\n" \
"MODULE_0=coretemp\n" \
"MODULE_1=f71882fg"
test: support empty environment variables in unit files Also update TODO, empty environment variables in Environment= and EnvironmentFile= options work.
2014-11-20 21:18:23 +01:00
#define env_file_5 \
"a=\n" \
"b="
fileio: in envfiles, do not skip lines following empty lines https://bugs.freedesktop.org/show_bug.cgi?id=63477
2013-04-14 02:22:53 +02:00
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
static void test_load_env_file_1(void) {
move _cleanup_ attribute in front of the type http://lists.freedesktop.org/archives/systemd-devel/2013-April/010510.html
2013-04-18 09:11:22 +02:00
_cleanup_strv_free_ char **data = NULL;
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
int r;
fs-util,test: add helper to remove tempfiles This simplifies the use of tempfiles in tests and fixes "leaked" temporary files in test-fileio, test-catalog, test-conf-parser. Not the whole tree is converted.
2018-05-16 11:35:41 +02:00
_cleanup_(unlink_tempfilep) char name[] = "/tmp/test-load-env-file.XXXXXX";
always use the same code for creating temporary files Let's unify our code here, and also always specifiy O_CLOEXEC.
2014-01-28 13:47:35 +01:00
_cleanup_close_ int fd;
fileio: simplify mkostemp_safe() (#4090) According to its manual page, flags given to mkostemp(3) shouldn't include O_RDWR, O_CREAT or O_EXCL flags as these are always included. Beyond those, the only flag that all callers (except a few tests where it probably doesn't matter) use is O_CLOEXEC, so set that unconditionally.
2016-09-13 08:20:38 +02:00
fd = mkostemp_safe(name);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(fd >= 0);
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
assert_se(write(fd, env_file_1, sizeof(env_file_1)) == sizeof(env_file_1));
fileio: drop "newline" parameter for env file parsers Now that we don't (mis-)use the env file parser to parse kernel command lines there's no need anymore to override the used newline character set. Let's hence drop the argument and just "\n\r" always. This nicely simplifies our code.
2018-11-12 14:04:47 +01:00
r = load_env_file(NULL, name, &data);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(r == 0);
assert_se(streq(data[0], "a=a"));
assert_se(streq(data[1], "b=bc"));
assert_se(streq(data[2], "d=def"));
assert_se(streq(data[3], "g=g "));
assert_se(streq(data[4], "h=h"));
assert_se(streq(data[5], "i=i"));
assert_se(data[6] == NULL);
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
}
static void test_load_env_file_2(void) {
move _cleanup_ attribute in front of the type http://lists.freedesktop.org/archives/systemd-devel/2013-April/010510.html
2013-04-18 09:11:22 +02:00
_cleanup_strv_free_ char **data = NULL;
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
int r;
fs-util,test: add helper to remove tempfiles This simplifies the use of tempfiles in tests and fixes "leaked" temporary files in test-fileio, test-catalog, test-conf-parser. Not the whole tree is converted.
2018-05-16 11:35:41 +02:00
_cleanup_(unlink_tempfilep) char name[] = "/tmp/test-load-env-file.XXXXXX";
always use the same code for creating temporary files Let's unify our code here, and also always specifiy O_CLOEXEC.
2014-01-28 13:47:35 +01:00
_cleanup_close_ int fd;
fileio: simplify mkostemp_safe() (#4090) According to its manual page, flags given to mkostemp(3) shouldn't include O_RDWR, O_CREAT or O_EXCL flags as these are always included. Beyond those, the only flag that all callers (except a few tests where it probably doesn't matter) use is O_CLOEXEC, so set that unconditionally.
2016-09-13 08:20:38 +02:00
fd = mkostemp_safe(name);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(fd >= 0);
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
assert_se(write(fd, env_file_2, sizeof(env_file_2)) == sizeof(env_file_2));
fileio: drop "newline" parameter for env file parsers Now that we don't (mis-)use the env file parser to parse kernel command lines there's no need anymore to override the used newline character set. Let's hence drop the argument and just "\n\r" always. This nicely simplifies our code.
2018-11-12 14:04:47 +01:00
r = load_env_file(NULL, name, &data);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(r == 0);
assert_se(streq(data[0], "a=a"));
assert_se(data[1] == NULL);
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
}
tests: add test for continuation lines with comments
2013-02-13 05:28:48 +01:00
static void test_load_env_file_3(void) {
move _cleanup_ attribute in front of the type http://lists.freedesktop.org/archives/systemd-devel/2013-April/010510.html
2013-04-18 09:11:22 +02:00
_cleanup_strv_free_ char **data = NULL;
tests: add test for continuation lines with comments
2013-02-13 05:28:48 +01:00
int r;
fs-util,test: add helper to remove tempfiles This simplifies the use of tempfiles in tests and fixes "leaked" temporary files in test-fileio, test-catalog, test-conf-parser. Not the whole tree is converted.
2018-05-16 11:35:41 +02:00
_cleanup_(unlink_tempfilep) char name[] = "/tmp/test-load-env-file.XXXXXX";
always use the same code for creating temporary files Let's unify our code here, and also always specifiy O_CLOEXEC.
2014-01-28 13:47:35 +01:00
_cleanup_close_ int fd;
fileio: simplify mkostemp_safe() (#4090) According to its manual page, flags given to mkostemp(3) shouldn't include O_RDWR, O_CREAT or O_EXCL flags as these are always included. Beyond those, the only flag that all callers (except a few tests where it probably doesn't matter) use is O_CLOEXEC, so set that unconditionally.
2016-09-13 08:20:38 +02:00
fd = mkostemp_safe(name);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(fd >= 0);
tests: add test for continuation lines with comments
2013-02-13 05:28:48 +01:00
assert_se(write(fd, env_file_3, sizeof(env_file_3)) == sizeof(env_file_3));
fileio: drop "newline" parameter for env file parsers Now that we don't (mis-)use the env file parser to parse kernel command lines there's no need anymore to override the used newline character set. Let's hence drop the argument and just "\n\r" always. This nicely simplifies our code.
2018-11-12 14:04:47 +01:00
r = load_env_file(NULL, name, &data);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(r == 0);
assert_se(data == NULL);
tests: add test for continuation lines with comments
2013-02-13 05:28:48 +01:00
}
fileio: in envfiles, do not skip lines following empty lines https://bugs.freedesktop.org/show_bug.cgi?id=63477
2013-04-14 02:22:53 +02:00
static void test_load_env_file_4(void) {
move _cleanup_ attribute in front of the type http://lists.freedesktop.org/archives/systemd-devel/2013-April/010510.html
2013-04-18 09:11:22 +02:00
_cleanup_strv_free_ char **data = NULL;
fs-util,test: add helper to remove tempfiles This simplifies the use of tempfiles in tests and fixes "leaked" temporary files in test-fileio, test-catalog, test-conf-parser. Not the whole tree is converted.
2018-05-16 11:35:41 +02:00
_cleanup_(unlink_tempfilep) char name[] = "/tmp/test-load-env-file.XXXXXX";
always use the same code for creating temporary files Let's unify our code here, and also always specifiy O_CLOEXEC.
2014-01-28 13:47:35 +01:00
_cleanup_close_ int fd;
fileio: in envfiles, do not skip lines following empty lines https://bugs.freedesktop.org/show_bug.cgi?id=63477
2013-04-14 02:22:53 +02:00
int r;
fileio: simplify mkostemp_safe() (#4090) According to its manual page, flags given to mkostemp(3) shouldn't include O_RDWR, O_CREAT or O_EXCL flags as these are always included. Beyond those, the only flag that all callers (except a few tests where it probably doesn't matter) use is O_CLOEXEC, so set that unconditionally.
2016-09-13 08:20:38 +02:00
fd = mkostemp_safe(name);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(fd >= 0);
fileio: in envfiles, do not skip lines following empty lines https://bugs.freedesktop.org/show_bug.cgi?id=63477
2013-04-14 02:22:53 +02:00
assert_se(write(fd, env_file_4, sizeof(env_file_4)) == sizeof(env_file_4));
fileio: drop "newline" parameter for env file parsers Now that we don't (mis-)use the env file parser to parse kernel command lines there's no need anymore to override the used newline character set. Let's hence drop the argument and just "\n\r" always. This nicely simplifies our code.
2018-11-12 14:04:47 +01:00
r = load_env_file(NULL, name, &data);
test: only use assert_se The asserts used in the tests should never be allowed to be optimized away
2014-10-04 23:51:45 +02:00
assert_se(r == 0);
assert_se(streq(data[0], "HWMON_MODULES=coretemp f71882fg"));
assert_se(streq(data[1], "MODULE_0=coretemp"));
assert_se(streq(data[2], "MODULE_1=f71882fg"));
assert_se(data[3] == NULL);
fileio: in envfiles, do not skip lines following empty lines https://bugs.freedesktop.org/show_bug.cgi?id=63477
2013-04-14 02:22:53 +02:00
}
test: support empty environment variables in unit files Also update TODO, empty environment variables in Environment= and EnvironmentFile= options work.
2014-11-20 21:18:23 +01:00
static void test_load_env_file_5(void) {
_cleanup_strv_free_ char **data = NULL;
int r;
fs-util,test: add helper to remove tempfiles This simplifies the use of tempfiles in tests and fixes "leaked" temporary files in test-fileio, test-catalog, test-conf-parser. Not the whole tree is converted.
2018-05-16 11:35:41 +02:00
_cleanup_(unlink_tempfilep) char name[] = "/tmp/test-load-env-file.XXXXXX";
test: support empty environment variables in unit files Also update TODO, empty environment variables in Environment= and EnvironmentFile= options work.
2014-11-20 21:18:23 +01:00
_cleanup_close_ int fd;
fileio: simplify mkostemp_safe() (#4090) According to its manual page, flags given to mkostemp(3) shouldn't include O_RDWR, O_CREAT or O_EXCL flags as these are always included. Beyond those, the only flag that all callers (except a few tests where it probably doesn't matter) use is O_CLOEXEC, so set that unconditionally.
2016-09-13 08:20:38 +02:00
fd = mkostemp_safe(name);
test: support empty environment variables in unit files Also update TODO, empty environment variables in Environment= and EnvironmentFile= options work.
2014-11-20 21:18:23 +01:00
assert_se(fd >= 0);
assert_se(write(fd, env_file_5, sizeof(env_file_5)) == sizeof(env_file_5));
fileio: drop "newline" parameter for env file parsers Now that we don't (mis-)use the env file parser to parse kernel command lines there's no need anymore to override the used newline character set. Let's hence drop the argument and just "\n\r" always. This nicely simplifies our code.
2018-11-12 14:04:47 +01:00
r = load_env_file(NULL, name, &data);
test: support empty environment variables in unit files Also update TODO, empty environment variables in Environment= and EnvironmentFile= options work.
2014-11-20 21:18:23 +01:00
assert_se(r == 0);
assert_se(streq(data[0], "a="));
assert_se(streq(data[1], "b="));
assert_se(data[2] == NULL);
}
fileio: in envfiles, do not skip lines following empty lines https://bugs.freedesktop.org/show_bug.cgi?id=63477
2013-04-14 02:22:53 +02:00
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
static void test_install_printf(void) {
char name[] = "name.service",
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
path[] = "/run/systemd/system/name.service";
UnitFileInstallInfo i = { .name = name, .path = path, };
UnitFileInstallInfo i2 = { .name= name, .path = path, };
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
char name3[] = "name@inst.service",
Fix typos of /etc/systemd/systemd <Lekensteyn> The 198 announcement mentions "/etc/systemd/systemd/foobar.service.d/*.conf", is that a typo? I only have a /etc/systemd/system/. Is there a manpage describing this new feature?
2013-03-10 00:55:31 +01:00
path3[] = "/run/systemd/system/name.service";
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
UnitFileInstallInfo i3 = { .name = name3, .path = path3, };
UnitFileInstallInfo i4 = { .name = name3, .path = path3, };
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
shared: add %g, %G specifiers for group / gid (#10368)
2018-10-13 10:26:48 +02:00
_cleanup_free_ char *mid = NULL, *bid = NULL, *host = NULL, *gid = NULL, *group = NULL, *uid = NULL, *user = NULL;
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
specifier: rework specifier calls to return proper error message Previously the specifier calls could only indicate OOM by returning NULL. With this change they will return negative errno-style error codes like everything else.
2013-09-17 17:03:46 +02:00
assert_se(specifier_machine_id('m', NULL, NULL, &mid) >= 0 && mid);
assert_se(specifier_boot_id('b', NULL, NULL, &bid) >= 0 && bid);
tree-wide: do not wrap assert_se in extra parentheses We were inconsitently using them in some cases, but in majority not. Using assignment in assert_se is very common, not an exception like in 'if', so let's drop the extra parens everywhere.
2018-05-14 09:12:12 +02:00
assert_se(host = gethostname_malloc());
shared: add %g, %G specifiers for group / gid (#10368)
2018-10-13 10:26:48 +02:00
assert_se(group = gid_to_name(getgid()));
assert_se(asprintf(&gid, UID_FMT, getgid()) >= 0);
tree-wide: do not wrap assert_se in extra parentheses We were inconsitently using them in some cases, but in majority not. Using assignment in assert_se is very common, not an exception like in 'if', so let's drop the extra parens everywhere.
2018-05-14 09:12:12 +02:00
assert_se(user = uid_to_name(getuid()));
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
assert_se(asprintf(&uid, UID_FMT, getuid()) >= 0);
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
#define expect(src, pattern, result) \
shared: rework env file reader Implement this with a proper state machine, so that newlines and escaped chars can appear in string assignments. This should bring the parser much closer to shell.
2013-04-03 19:04:03 +02:00
do { \
specifier: rework specifier calls to return proper error message Previously the specifier calls could only indicate OOM by returning NULL. With this change they will return negative errno-style error codes like everything else.
2013-09-17 17:03:46 +02:00
_cleanup_free_ char *t = NULL; \
move _cleanup_ attribute in front of the type http://lists.freedesktop.org/archives/systemd-devel/2013-April/010510.html
2013-04-18 09:11:22 +02:00
_cleanup_free_ char \
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
*d1 = strdup(i.name), \
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
*d2 = strdup(i.path); \
specifier: rework specifier calls to return proper error message Previously the specifier calls could only indicate OOM by returning NULL. With this change they will return negative errno-style error codes like everything else.
2013-09-17 17:03:46 +02:00
assert_se(install_full_printf(&src, pattern, &t) >= 0 || !result); \
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
memzero(i.name, strlen(i.name)); \
memzero(i.path, strlen(i.path)); \
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
assert_se(d1 && d2); \
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
if (result) { \
printf("%s\n", t); \
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
assert_se(streq(t, result)); \
} else assert_se(t == NULL); \
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
strcpy(i.name, d1); \
strcpy(i.path, d2); \
tree-wide: minor formatting inconsistency cleanups
2016-02-23 18:52:52 +01:00
} while (false)
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
expect(i, "%n", "name.service");
expect(i, "%N", "name");
expect(i, "%p", "name");
expect(i, "%i", "");
Add %j/%J unit specifiers Those are quite similar to %i/%I, but refer to the last dash-separated component of the name prefix. The new functionality of dash-dropins could largely supersede the template functionality, so it would be tempting to overload %i/%I. But that would not be backwards compatible. So let's add the two new letters instead.
2018-04-23 22:43:20 +02:00
expect(i, "%j", "name");
shared: add %g, %G specifiers for group / gid (#10368)
2018-10-13 10:26:48 +02:00
expect(i, "%g", group);
expect(i, "%G", gid);
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
expect(i, "%u", user);
expect(i, "%U", uid);
macro: add nice macro for disabling -Wnonnull temporarily
2014-02-20 18:44:51 +01:00
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
expect(i, "%m", mid);
expect(i, "%b", bid);
expect(i, "%H", host);
shared: add %g, %G specifiers for group / gid (#10368)
2018-10-13 10:26:48 +02:00
expect(i2, "%g", group);
expect(i2, "%G", gid);
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
expect(i2, "%u", user);
expect(i2, "%U", uid);
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
expect(i3, "%n", "name@inst.service");
expect(i3, "%N", "name@inst");
expect(i3, "%p", "name");
shared: add %g, %G specifiers for group / gid (#10368)
2018-10-13 10:26:48 +02:00
expect(i3, "%g", group);
expect(i3, "%G", gid);
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
expect(i3, "%u", user);
expect(i3, "%U", uid);
macro: add nice macro for disabling -Wnonnull temporarily
2014-02-20 18:44:51 +01:00
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
expect(i3, "%m", mid);
expect(i3, "%b", bid);
expect(i3, "%H", host);
shared: add %g, %G specifiers for group / gid (#10368)
2018-10-13 10:26:48 +02:00
expect(i4, "%g", group);
expect(i4, "%G", gid);
core: simplify handling of %u, %U, %s and %h unit file specifiers Previously, the %u, %U, %s and %h specifiers would resolve to the user name, numeric user ID, shell and home directory of the user configured in the User= setting of a unit file, or the user of the manager instance if no User= setting was configured. That at least was the theory. In real-life this was not ever actually useful: - For the systemd --user instance it made no sense to ever set User=, since the instance runs in user context after all, and hence the privileges to change user IDs don't even exist. The four specifiers were actually not useful at all in this case. - For the systemd --system instance we did not allow any resolving that would require NSS. Hence, %s and %h were not supported, unless User=root was set, in which case they would be hardcoded to /bin/sh and /root, to avoid NSS. Then, %u would actually resolve to whatever was set with User=, but %U would only resolve to the numeric UID of that setting if the User= was specified in numeric form, or happened to be root (in which case 0 was hardcoded as mapping). Two of the specifiers are entirely useless in this case, one is realistically also useless, and one is pretty pointless. - Resolving of these settings would only happen if User= was actually set *before* the specifiers where resolved. This behaviour was undocumented and is really ugly, as specifiers should actually be considered something that applies to the whole file equally, independently of order... With this change, %u, %U, %s and %h are drastically simplified: they now always refer to the user that is running the service instance, and the user configured in the unit file is irrelevant. For the system instance of systemd this means they always resolve to "root", "0", "/bin/sh" and "/root", thus avoiding NSS. For the user instance, to the data for the specific user. The new behaviour is identical to the old behaviour in all --user cases and for all units that have no User= set (or set to "0" or "root").
2015-10-31 22:12:51 +01:00
expect(i4, "%u", user);
expect(i4, "%U", uid);
tests: add test for install_full_printf()
2013-01-29 20:39:06 +01:00
}
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
static uint64_t make_cap(int cap) {
return ((uint64_t) 1ULL << (uint64_t) cap);
}
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
static void test_config_parse_capability_set(void) {
/* int config_parse_capability_set(
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
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 r;
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
uint64_t capability_bounding_set = 0;
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
r = config_parse_capability_set(NULL, "fake", 1, "section", 1,
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
"CapabilityBoundingSet", 0, "CAP_NET_RAW",
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
&capability_bounding_set, NULL);
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
assert_se(r >= 0);
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
assert_se(capability_bounding_set == make_cap(CAP_NET_RAW));
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
r = config_parse_capability_set(NULL, "fake", 1, "section", 1,
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
"CapabilityBoundingSet", 0, "CAP_NET_ADMIN",
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
&capability_bounding_set, NULL);
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
assert_se(r >= 0);
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
assert_se(capability_bounding_set == (make_cap(CAP_NET_RAW) | make_cap(CAP_NET_ADMIN)));
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
test: update test-unit-file Follow-up for c792ec2e3512a672881fc847ff432e26b641c9c9.
2017-09-04 09:25:22 +02:00
r = config_parse_capability_set(NULL, "fake", 1, "section", 1,
"CapabilityBoundingSet", 0, "~CAP_NET_ADMIN",
&capability_bounding_set, NULL);
assert_se(r >= 0);
assert_se(capability_bounding_set == make_cap(CAP_NET_RAW));
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
r = config_parse_capability_set(NULL, "fake", 1, "section", 1,
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
"CapabilityBoundingSet", 0, "",
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
&capability_bounding_set, NULL);
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
assert_se(r >= 0);
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
assert_se(capability_bounding_set == UINT64_C(0));
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
r = config_parse_capability_set(NULL, "fake", 1, "section", 1,
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
"CapabilityBoundingSet", 0, "~",
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
&capability_bounding_set, NULL);
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
assert_se(r >= 0);
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
assert_se(cap_test_all(capability_bounding_set));
tests: add test for capability bounding set parsing
2015-10-30 07:22:53 +01:00
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
capability_bounding_set = 0;
r = config_parse_capability_set(NULL, "fake", 1, "section", 1,
tests: add test for capability bounding set parsing
2015-10-30 07:22:53 +01:00
"CapabilityBoundingSet", 0, " 'CAP_NET_RAW' WAT_CAP??? CAP_NET_ADMIN CAP'_trailing_garbage",
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
&capability_bounding_set, NULL);
tests: add test for capability bounding set parsing
2015-10-30 07:22:53 +01:00
assert_se(r >= 0);
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
assert_se(capability_bounding_set == (make_cap(CAP_NET_RAW) | make_cap(CAP_NET_ADMIN)));
test: add test for capability bounding set parsing
2015-10-29 12:12:22 +01:00
}
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
static void test_config_parse_rlimit(void) {
struct rlimit * rl[_RLIMIT_MAX] = {};
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitNOFILE", RLIMIT_NOFILE, "55", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_NOFILE]);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == 55);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == rl[RLIMIT_NOFILE]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitNOFILE", RLIMIT_NOFILE, "55:66", rl, NULL) >= 0);
core: support <soft:hard> ranges for RLIMIT options The new parser supports: <value> - specify both limits to the same value <soft:hard> - specify both limits the size or time specific suffixes are supported, for example LimitRTTIME=1sec LimitAS=4G:16G The patch introduces parse_rlimit_range() and rlim type (size, sec, usec, etc.) specific parsers. No code is duplicated now. The patch also sync docs for DefaultLimitXXX= and LimitXXX=. References: https://github.com/systemd/systemd/issues/1769
2015-11-20 12:54:10 +01:00
assert_se(rl[RLIMIT_NOFILE]);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == 55);
assert_se(rl[RLIMIT_NOFILE]->rlim_max == 66);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitNOFILE", RLIMIT_NOFILE, "infinity", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_NOFILE]);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == RLIM_INFINITY);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == rl[RLIMIT_NOFILE]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitNOFILE", RLIMIT_NOFILE, "infinity:infinity", rl, NULL) >= 0);
core: support <soft:hard> ranges for RLIMIT options The new parser supports: <value> - specify both limits to the same value <soft:hard> - specify both limits the size or time specific suffixes are supported, for example LimitRTTIME=1sec LimitAS=4G:16G The patch introduces parse_rlimit_range() and rlim type (size, sec, usec, etc.) specific parsers. No code is duplicated now. The patch also sync docs for DefaultLimitXXX= and LimitXXX=. References: https://github.com/systemd/systemd/issues/1769
2015-11-20 12:54:10 +01:00
assert_se(rl[RLIMIT_NOFILE]);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == RLIM_INFINITY);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == rl[RLIMIT_NOFILE]->rlim_max);
core: move parsing of rlimits into rlimit-util.[ch] This way we can reuse it for parsing rlimit settings in "systemctl set-property" and related commands.
2016-02-01 21:07:09 +01:00
rl[RLIMIT_NOFILE]->rlim_cur = 10;
rl[RLIMIT_NOFILE]->rlim_max = 20;
/* Invalid values don't change rl */
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitNOFILE", RLIMIT_NOFILE, "10:20:30", rl, NULL) >= 0);
core: fix rlimit parsing * refuse limits if soft > hard * print an actual value instead of (null) see https://github.com/systemd/systemd/pull/1994#issuecomment-159999123
2015-11-27 09:54:42 +01:00
assert_se(rl[RLIMIT_NOFILE]);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == 10);
assert_se(rl[RLIMIT_NOFILE]->rlim_max == 20);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitNOFILE", RLIMIT_NOFILE, "wat:wat", rl, NULL) >= 0);
core: fix rlimit parsing * refuse limits if soft > hard * print an actual value instead of (null) see https://github.com/systemd/systemd/pull/1994#issuecomment-159999123
2015-11-27 09:54:42 +01:00
assert_se(rl[RLIMIT_NOFILE]);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == 10);
assert_se(rl[RLIMIT_NOFILE]->rlim_max == 20);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitNOFILE", RLIMIT_NOFILE, "66:wat", rl, NULL) >= 0);
core: fix rlimit parsing * refuse limits if soft > hard * print an actual value instead of (null) see https://github.com/systemd/systemd/pull/1994#issuecomment-159999123
2015-11-27 09:54:42 +01:00
assert_se(rl[RLIMIT_NOFILE]);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == 10);
assert_se(rl[RLIMIT_NOFILE]->rlim_max == 20);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitNOFILE", RLIMIT_NOFILE, "200:100", rl, NULL) >= 0);
core: fix rlimit parsing * refuse limits if soft > hard * print an actual value instead of (null) see https://github.com/systemd/systemd/pull/1994#issuecomment-159999123
2015-11-27 09:54:42 +01:00
assert_se(rl[RLIMIT_NOFILE]);
assert_se(rl[RLIMIT_NOFILE]->rlim_cur == 10);
assert_se(rl[RLIMIT_NOFILE]->rlim_max == 20);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
rl[RLIMIT_NOFILE] = mfree(rl[RLIMIT_NOFILE]);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitCPU", RLIMIT_CPU, "56", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_CPU]);
assert_se(rl[RLIMIT_CPU]->rlim_cur == 56);
assert_se(rl[RLIMIT_CPU]->rlim_cur == rl[RLIMIT_CPU]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitCPU", RLIMIT_CPU, "57s", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_CPU]);
assert_se(rl[RLIMIT_CPU]->rlim_cur == 57);
assert_se(rl[RLIMIT_CPU]->rlim_cur == rl[RLIMIT_CPU]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitCPU", RLIMIT_CPU, "40s:1m", rl, NULL) >= 0);
core: support <soft:hard> ranges for RLIMIT options The new parser supports: <value> - specify both limits to the same value <soft:hard> - specify both limits the size or time specific suffixes are supported, for example LimitRTTIME=1sec LimitAS=4G:16G The patch introduces parse_rlimit_range() and rlim type (size, sec, usec, etc.) specific parsers. No code is duplicated now. The patch also sync docs for DefaultLimitXXX= and LimitXXX=. References: https://github.com/systemd/systemd/issues/1769
2015-11-20 12:54:10 +01:00
assert_se(rl[RLIMIT_CPU]);
assert_se(rl[RLIMIT_CPU]->rlim_cur == 40);
assert_se(rl[RLIMIT_CPU]->rlim_max == 60);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitCPU", RLIMIT_CPU, "infinity", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_CPU]);
assert_se(rl[RLIMIT_CPU]->rlim_cur == RLIM_INFINITY);
assert_se(rl[RLIMIT_CPU]->rlim_cur == rl[RLIMIT_CPU]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitCPU", RLIMIT_CPU, "1234ms", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_CPU]);
assert_se(rl[RLIMIT_CPU]->rlim_cur == 2);
assert_se(rl[RLIMIT_CPU]->rlim_cur == rl[RLIMIT_CPU]->rlim_max);
rl[RLIMIT_CPU] = mfree(rl[RLIMIT_CPU]);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitRTTIME", RLIMIT_RTTIME, "58", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_RTTIME]);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == 58);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == rl[RLIMIT_RTTIME]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitRTTIME", RLIMIT_RTTIME, "58:60", rl, NULL) >= 0);
core: support <soft:hard> ranges for RLIMIT options The new parser supports: <value> - specify both limits to the same value <soft:hard> - specify both limits the size or time specific suffixes are supported, for example LimitRTTIME=1sec LimitAS=4G:16G The patch introduces parse_rlimit_range() and rlim type (size, sec, usec, etc.) specific parsers. No code is duplicated now. The patch also sync docs for DefaultLimitXXX= and LimitXXX=. References: https://github.com/systemd/systemd/issues/1769
2015-11-20 12:54:10 +01:00
assert_se(rl[RLIMIT_RTTIME]);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == 58);
assert_se(rl[RLIMIT_RTTIME]->rlim_max == 60);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitRTTIME", RLIMIT_RTTIME, "59s", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_RTTIME]);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == 59 * USEC_PER_SEC);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == rl[RLIMIT_RTTIME]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitRTTIME", RLIMIT_RTTIME, "59s:123s", rl, NULL) >= 0);
core: support <soft:hard> ranges for RLIMIT options The new parser supports: <value> - specify both limits to the same value <soft:hard> - specify both limits the size or time specific suffixes are supported, for example LimitRTTIME=1sec LimitAS=4G:16G The patch introduces parse_rlimit_range() and rlim type (size, sec, usec, etc.) specific parsers. No code is duplicated now. The patch also sync docs for DefaultLimitXXX= and LimitXXX=. References: https://github.com/systemd/systemd/issues/1769
2015-11-20 12:54:10 +01:00
assert_se(rl[RLIMIT_RTTIME]);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == 59 * USEC_PER_SEC);
assert_se(rl[RLIMIT_RTTIME]->rlim_max == 123 * USEC_PER_SEC);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitRTTIME", RLIMIT_RTTIME, "infinity", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_RTTIME]);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == RLIM_INFINITY);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == rl[RLIMIT_RTTIME]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitRTTIME", RLIMIT_RTTIME, "infinity:infinity", rl, NULL) >= 0);
core: support <soft:hard> ranges for RLIMIT options The new parser supports: <value> - specify both limits to the same value <soft:hard> - specify both limits the size or time specific suffixes are supported, for example LimitRTTIME=1sec LimitAS=4G:16G The patch introduces parse_rlimit_range() and rlim type (size, sec, usec, etc.) specific parsers. No code is duplicated now. The patch also sync docs for DefaultLimitXXX= and LimitXXX=. References: https://github.com/systemd/systemd/issues/1769
2015-11-20 12:54:10 +01:00
assert_se(rl[RLIMIT_RTTIME]);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == RLIM_INFINITY);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == rl[RLIMIT_RTTIME]->rlim_max);
core: move config_parse_limit() to the generic conf-parser.[ch] That way we can use it in nspawn. Also, while we are at it, let's rename the call config_parse_rlimit(), i.e. insert the "r", to clarify what kind of limit this is about.
2018-05-03 19:01:21 +02:00
assert_se(config_parse_rlimit(NULL, "fake", 1, "section", 1, "LimitRTTIME", RLIMIT_RTTIME, "2345ms", rl, NULL) >= 0);
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
assert_se(rl[RLIMIT_RTTIME]);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == 2345 * USEC_PER_MSEC);
assert_se(rl[RLIMIT_RTTIME]->rlim_cur == rl[RLIMIT_RTTIME]->rlim_max);
rl[RLIMIT_RTTIME] = mfree(rl[RLIMIT_RTTIME]);
}
tests: add test for config_parse_pass_environ
2015-11-07 19:54:12 +01:00
static void test_config_parse_pass_environ(void) {
/* int config_parse_pass_environ(
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 r;
_cleanup_strv_free_ char **passenv = NULL;
r = config_parse_pass_environ(NULL, "fake", 1, "section", 1,
"PassEnvironment", 0, "A B",
&passenv, NULL);
assert_se(r >= 0);
assert_se(strv_length(passenv) == 2);
assert_se(streq(passenv[0], "A"));
assert_se(streq(passenv[1], "B"));
r = config_parse_pass_environ(NULL, "fake", 1, "section", 1,
"PassEnvironment", 0, "",
&passenv, NULL);
assert_se(r >= 0);
assert_se(strv_isempty(passenv));
r = config_parse_pass_environ(NULL, "fake", 1, "section", 1,
"PassEnvironment", 0, "'invalid name' 'normal_name' A=1 \\",
&passenv, NULL);
assert_se(r >= 0);
assert_se(strv_length(passenv) == 1);
assert_se(streq(passenv[0], "normal_name"));
}
build-sys: fix the script used for generating load_fragment_gperf_nulstr (#6646) This prevents `systemd --dump-configuration-items` from crashing. Closes #6596.
2017-08-26 16:21:23 +02:00
static void test_unit_dump_config_items(void) {
unit_dump_config_items(stdout);
}
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
int main(int argc, char *argv[]) {
basic: remove rm_rf_and_free, add rm_rf_physical_and_free, use rm_rf_physical_and_freep in tests (#3292) Some distros don't mount /tmp as tmpfs. For example: https://lists.ubuntu.com/archives/ubuntu-cloud/2016-January/001009.html Some tests: * print 'Attempted to remove disk file system, and we can't allow that.' * don't really cleanup /tmp
2016-05-20 15:08:24 +02:00
_cleanup_(rm_rf_physical_and_freep) char *runtime_dir = NULL;
test-unit-file: return error without dumping core on permission error
2013-07-30 15:07:20 +02:00
int r;
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
tests: use a helper function to parse environment and open logging The advantages are that we save a few lines, and that we can override logging using environment variables in more test executables.
2018-09-13 14:31:13 +02:00
test_setup_logging(LOG_INFO);
core: log a few more things under UNIT=...
2013-04-15 07:40:44 +02:00
tests: skip tests when cg_pid_get_path fails (#7033) v2: - cast the fstype_t type to ull, because it varies between arches. Making it long long should be on the safe side.
2017-10-10 20:55:20 +02:00
r = enter_cgroup_subroot();
tests: add helper to unify skipping a test and exiting
2018-09-13 13:34:12 +02:00
if (r == -ENOMEDIUM)
return log_tests_skipped("cgroupfs not available");
tests: when running a manager object in a test, migrate to private cgroup subroot first (#6576) Without this "meson test" will end up running all tests in the same cgroup root, and they all will try to manage it. Which usually isn't too bad, except when they end up clearing up each other's cgroups. This race is hard to trigger but has caused various CI runs to fail spuriously. With this change we simply move every test that runs a manager object into their own private cgroup. Note that we don't clean up the cgroup at the end, we leave that to the cgroup manager around it. This fixes races that become visible by test runs throwing out errors like this: ``` exec-systemcallfilter-failing.service: Passing 0 fds to service exec-systemcallfilter-failing.service: About to execute: /bin/echo 'This should not be seen' exec-systemcallfilter-failing.service: Forked /bin/echo as 5693 exec-systemcallfilter-failing.service: Changed dead -> start exec-systemcallfilter-failing.service: Failed to attach to cgroup /exec-systemcallfilter-failing.service: No such file or directory Received SIGCHLD from PID 5693 ((echo)). Child 5693 ((echo)) died (code=exited, status=219/CGROUP) exec-systemcallfilter-failing.service: Child 5693 belongs to exec-systemcallfilter-failing.service exec-systemcallfilter-failing.service: Main process exited, code=exited, status=219/CGROUP exec-systemcallfilter-failing.service: Changed start -> failed exec-systemcallfilter-failing.service: Unit entered failed state. exec-systemcallfilter-failing.service: Failed with result 'exit-code'. exec-systemcallfilter-failing.service: cgroup is empty Assertion 'service->main_exec_status.status == status_expected' failed at ../src/src/test/test-execute.c:71, function check(). Aborting. ``` BTW, I tracked this race down by using perf: ``` # perf record -e cgroup:cgroup_mkdir,cgroup_rmdir … # perf script ``` Thanks a lot @iaguis, @alban for helping me how to use perf for this. Fixes #5895.
2017-08-09 15:42:49 +02:00
tests: override XDG_RUNTIME_DIR where we use the user runtime dir We don#t really support systems where XDG_RUNTIME_DIR is not supported for systemd --user. Hence, let's always set our own XDG_RUNTIME_DIR for tests that involve systemd --user, so that we know it is set, and that it doesn't polute the user's actual runtime dir.
2016-04-08 18:54:05 +02:00
assert_se(runtime_dir = setup_fake_runtime_dir());
test-unit-file: return error without dumping core on permission error
2013-07-30 15:07:20 +02:00
r = test_unit_file_get_set();
tests: add tests for config_parse_exec
2012-11-15 14:48:12 +01:00
test_config_parse_exec();
test-unit-file: add test for config_parse_log_extra_fields()
2017-11-14 14:58:37 +01:00
test_config_parse_log_extra_fields();
capabilities: keep bounding set in non-inverted format. Change the capability bounding set parser and logic so that the bounding set is kept as a positive set internally. This means that the set reflects those capabilities that we want to keep instead of drop.
2016-01-07 23:00:04 +01:00
test_config_parse_capability_set();
core: accept time units for time-based resource limits Let's make sure "LimitCPU=30min" can be parsed properly, following the usual logic how we parse time values. Similar for LimitRTTIME=. While we are at it, extend a bit on the man page section about resource limits. Fixes: #1772
2015-11-10 16:52:52 +01:00
test_config_parse_rlimit();
tests: add test for config_parse_pass_environ
2015-11-07 19:54:12 +01:00
test_config_parse_pass_environ();
util: modernization and test for load_env_file
2013-01-18 16:57:11 +01:00
test_load_env_file_1();
test_load_env_file_2();
tests: add test for continuation lines with comments
2013-02-13 05:28:48 +01:00
test_load_env_file_3();
fileio: in envfiles, do not skip lines following empty lines https://bugs.freedesktop.org/show_bug.cgi?id=63477
2013-04-14 02:22:53 +02:00
test_load_env_file_4();
test: support empty environment variables in unit files Also update TODO, empty environment variables in Environment= and EnvironmentFile= options work.
2014-11-20 21:18:23 +01:00
test_load_env_file_5();
test: Make testing work on systems without or old systemd * Introduce a macro to conditionally execute tests. This avoids skipping the entire test if some parts require systemd * Skip the journal tests when no /etc/machine-id is present * Change test-catalog to load the catalog from the source directory of systemd. * /proc/PID/comm got introduced in v2.6.33 but travis is still using v2.6.32. * Enable make check and make distcheck on the travis build * Use -D"CATALOG_DIR=STR($(abs_top_srcdir)/catalog)" as a STRINGIY would result in the path '/home/ich/source/linux' to be expanded to '/home/ich/source/1' as linux is defined to 1.
2013-07-18 08:30:06 +02:00
TEST_REQ_RUNNING_SYSTEMD(test_install_printf());
build-sys: fix the script used for generating load_fragment_gperf_nulstr (#6646) This prevents `systemd --dump-configuration-items` from crashing. Closes #6596.
2017-08-26 16:21:23 +02:00
test_unit_dump_config_items();
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
test-unit-file: return error without dumping core on permission error
2013-07-30 15:07:20 +02:00
return r;
when determining unit file list, include invalid unit names in an "invalid" state
2012-09-11 01:11:32 +02:00
}
Reference in a new issue Copy permalink