summaryrefslogtreecommitdiff
path: root/src/core
diff options
context:
space:
mode:
Diffstat (limited to 'src/core')
-rw-r--r--src/core/cgroup.c185
-rw-r--r--src/core/cgroup.h1
-rw-r--r--src/core/dbus-execute.c5
-rw-r--r--src/core/execute.c53
-rw-r--r--src/core/execute.h1
-rw-r--r--src/core/load-fragment-gperf.gperf.m42
-rw-r--r--src/core/load-fragment.c99
7 files changed, 237 insertions, 109 deletions
diff --git a/src/core/cgroup.c b/src/core/cgroup.c
index fbe69df4e9..f3e0c54b76 100644
--- a/src/core/cgroup.c
+++ b/src/core/cgroup.c
@@ -36,6 +36,22 @@
#define CGROUP_CPU_QUOTA_PERIOD_USEC ((usec_t) 100 * USEC_PER_MSEC)
+static void cgroup_compat_warn(void)
+{
+ static bool cgroup_compat_warned = false;
+
+ if (cgroup_compat_warned)
+ return;
+
+ log_warning("cgroup compatibility translation between legacy and unified hierarchy settings activated. See cgroup-compat debug messages for details.");
+ cgroup_compat_warned = true;
+}
+
+#define log_cgroup_compat(unit, fmt, ...) do { \
+ cgroup_compat_warn(); \
+ log_unit_debug(unit, "cgroup-compat: " fmt, ##__VA_ARGS__); \
+ } while (0)
+
void cgroup_context_init(CGroupContext *c) {
assert(c);
@@ -413,7 +429,7 @@ static uint64_t cgroup_weight_io_to_blkio(uint64_t io_weight) {
CGROUP_BLKIO_WEIGHT_MIN, CGROUP_BLKIO_WEIGHT_MAX);
}
-static void cgroup_apply_io_device_weight(const char *path, const char *dev_path, uint64_t io_weight) {
+static void cgroup_apply_io_device_weight(Unit *u, const char *dev_path, uint64_t io_weight) {
char buf[DECIMAL_STR_MAX(dev_t)*2+2+DECIMAL_STR_MAX(uint64_t)+1];
dev_t dev;
int r;
@@ -423,13 +439,13 @@ static void cgroup_apply_io_device_weight(const char *path, const char *dev_path
return;
xsprintf(buf, "%u:%u %" PRIu64 "\n", major(dev), minor(dev), io_weight);
- r = cg_set_attribute("io", path, "io.weight", buf);
+ r = cg_set_attribute("io", u->cgroup_path, "io.weight", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set io.weight on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set io.weight: %m");
}
-static void cgroup_apply_blkio_device_weight(const char *path, const char *dev_path, uint64_t blkio_weight) {
+static void cgroup_apply_blkio_device_weight(Unit *u, const char *dev_path, uint64_t blkio_weight) {
char buf[DECIMAL_STR_MAX(dev_t)*2+2+DECIMAL_STR_MAX(uint64_t)+1];
dev_t dev;
int r;
@@ -439,13 +455,13 @@ static void cgroup_apply_blkio_device_weight(const char *path, const char *dev_p
return;
xsprintf(buf, "%u:%u %" PRIu64 "\n", major(dev), minor(dev), blkio_weight);
- r = cg_set_attribute("blkio", path, "blkio.weight_device", buf);
+ r = cg_set_attribute("blkio", u->cgroup_path, "blkio.weight_device", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set blkio.weight_device on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set blkio.weight_device: %m");
}
-static unsigned cgroup_apply_io_device_limit(const char *path, const char *dev_path, uint64_t *limits) {
+static unsigned cgroup_apply_io_device_limit(Unit *u, const char *dev_path, uint64_t *limits) {
char limit_bufs[_CGROUP_IO_LIMIT_TYPE_MAX][DECIMAL_STR_MAX(uint64_t)];
char buf[DECIMAL_STR_MAX(dev_t)*2+2+(6+DECIMAL_STR_MAX(uint64_t)+1)*4];
CGroupIOLimitType type;
@@ -469,14 +485,14 @@ static unsigned cgroup_apply_io_device_limit(const char *path, const char *dev_p
xsprintf(buf, "%u:%u rbps=%s wbps=%s riops=%s wiops=%s\n", major(dev), minor(dev),
limit_bufs[CGROUP_IO_RBPS_MAX], limit_bufs[CGROUP_IO_WBPS_MAX],
limit_bufs[CGROUP_IO_RIOPS_MAX], limit_bufs[CGROUP_IO_WIOPS_MAX]);
- r = cg_set_attribute("io", path, "io.max", buf);
+ r = cg_set_attribute("io", u->cgroup_path, "io.max", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set io.max on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set io.max: %m");
return n;
}
-static unsigned cgroup_apply_blkio_device_limit(const char *path, const char *dev_path, uint64_t rbps, uint64_t wbps) {
+static unsigned cgroup_apply_blkio_device_limit(Unit *u, const char *dev_path, uint64_t rbps, uint64_t wbps) {
char buf[DECIMAL_STR_MAX(dev_t)*2+2+DECIMAL_STR_MAX(uint64_t)+1];
dev_t dev;
unsigned n = 0;
@@ -489,18 +505,18 @@ static unsigned cgroup_apply_blkio_device_limit(const char *path, const char *de
if (rbps != CGROUP_LIMIT_MAX)
n++;
sprintf(buf, "%u:%u %" PRIu64 "\n", major(dev), minor(dev), rbps);
- r = cg_set_attribute("blkio", path, "blkio.throttle.read_bps_device", buf);
+ r = cg_set_attribute("blkio", u->cgroup_path, "blkio.throttle.read_bps_device", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set blkio.throttle.read_bps_device on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set blkio.throttle.read_bps_device: %m");
if (wbps != CGROUP_LIMIT_MAX)
n++;
sprintf(buf, "%u:%u %" PRIu64 "\n", major(dev), minor(dev), wbps);
- r = cg_set_attribute("blkio", path, "blkio.throttle.write_bps_device", buf);
+ r = cg_set_attribute("blkio", u->cgroup_path, "blkio.throttle.write_bps_device", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set blkio.throttle.write_bps_device on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set blkio.throttle.write_bps_device: %m");
return n;
}
@@ -509,23 +525,30 @@ static bool cgroup_context_has_unified_memory_config(CGroupContext *c) {
return c->memory_low > 0 || c->memory_high != CGROUP_LIMIT_MAX || c->memory_max != CGROUP_LIMIT_MAX;
}
-static void cgroup_apply_unified_memory_limit(const char *path, const char *file, uint64_t v) {
+static void cgroup_apply_unified_memory_limit(Unit *u, const char *file, uint64_t v) {
char buf[DECIMAL_STR_MAX(uint64_t) + 1] = "max";
int r;
if (v != CGROUP_LIMIT_MAX)
xsprintf(buf, "%" PRIu64 "\n", v);
- r = cg_set_attribute("memory", path, file, buf);
+ r = cg_set_attribute("memory", u->cgroup_path, file, buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set %s on %s: %m", file, path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set %s: %m", file);
}
-void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, ManagerState state) {
+static void cgroup_context_apply(Unit *u, CGroupMask mask, ManagerState state) {
+ const char *path;
+ CGroupContext *c;
bool is_root;
int r;
+ assert(u);
+
+ c = unit_get_cgroup_context(u);
+ path = u->cgroup_path;
+
assert(c);
assert(path);
@@ -551,14 +574,14 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
c->cpu_shares != CGROUP_CPU_SHARES_INVALID ? c->cpu_shares : CGROUP_CPU_SHARES_DEFAULT);
r = cg_set_attribute("cpu", path, "cpu.shares", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set cpu.shares on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set cpu.shares: %m");
sprintf(buf, USEC_FMT "\n", CGROUP_CPU_QUOTA_PERIOD_USEC);
r = cg_set_attribute("cpu", path, "cpu.cfs_period_us", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set cpu.cfs_period_us on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set cpu.cfs_period_us: %m");
if (c->cpu_quota_per_sec_usec != USEC_INFINITY) {
sprintf(buf, USEC_FMT "\n", c->cpu_quota_per_sec_usec * CGROUP_CPU_QUOTA_PERIOD_USEC / USEC_PER_SEC);
@@ -566,8 +589,8 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
} else
r = cg_set_attribute("cpu", path, "cpu.cfs_quota_us", "-1");
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set cpu.cfs_quota_us on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set cpu.cfs_quota_us: %m");
}
if (mask & CGROUP_MASK_IO) {
@@ -580,29 +603,40 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
if (has_io)
weight = cgroup_context_io_weight(c, state);
- else if (has_blockio)
- weight = cgroup_weight_blkio_to_io(cgroup_context_blkio_weight(c, state));
- else
+ else if (has_blockio) {
+ uint64_t blkio_weight = cgroup_context_blkio_weight(c, state);
+
+ weight = cgroup_weight_blkio_to_io(blkio_weight);
+
+ log_cgroup_compat(u, "Applying [Startup]BlockIOWeight %" PRIu64 " as [Startup]IOWeight %" PRIu64,
+ blkio_weight, weight);
+ } else
weight = CGROUP_WEIGHT_DEFAULT;
xsprintf(buf, "default %" PRIu64 "\n", weight);
r = cg_set_attribute("io", path, "io.weight", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set io.weight on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set io.weight: %m");
if (has_io) {
CGroupIODeviceWeight *w;
/* FIXME: no way to reset this list */
LIST_FOREACH(device_weights, w, c->io_device_weights)
- cgroup_apply_io_device_weight(path, w->path, w->weight);
+ cgroup_apply_io_device_weight(u, w->path, w->weight);
} else if (has_blockio) {
CGroupBlockIODeviceWeight *w;
/* FIXME: no way to reset this list */
- LIST_FOREACH(device_weights, w, c->blockio_device_weights)
- cgroup_apply_io_device_weight(path, w->path, cgroup_weight_blkio_to_io(w->weight));
+ LIST_FOREACH(device_weights, w, c->blockio_device_weights) {
+ weight = cgroup_weight_blkio_to_io(w->weight);
+
+ log_cgroup_compat(u, "Applying BlockIODeviceWeight %" PRIu64 " as IODeviceWeight %" PRIu64 " for %s",
+ w->weight, weight, w->path);
+
+ cgroup_apply_io_device_weight(u, w->path, weight);
+ }
}
}
@@ -611,7 +645,7 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
CGroupIODeviceLimit *l, *next;
LIST_FOREACH_SAFE(device_limits, l, next, c->io_device_limits) {
- if (!cgroup_apply_io_device_limit(path, l->path, l->limits))
+ if (!cgroup_apply_io_device_limit(u, l->path, l->limits))
cgroup_context_free_io_device_limit(c, l);
}
} else if (has_blockio) {
@@ -627,7 +661,10 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
limits[CGROUP_IO_RBPS_MAX] = b->rbps;
limits[CGROUP_IO_WBPS_MAX] = b->wbps;
- if (!cgroup_apply_io_device_limit(path, b->path, limits))
+ log_cgroup_compat(u, "Applying BlockIO{Read|Write}Bandwidth %" PRIu64 " %" PRIu64 " as IO{Read|Write}BandwidthMax for %s",
+ b->rbps, b->wbps, b->path);
+
+ if (!cgroup_apply_io_device_limit(u, b->path, limits))
cgroup_context_free_blockio_device_bandwidth(c, b);
}
}
@@ -643,29 +680,40 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
if (has_blockio)
weight = cgroup_context_blkio_weight(c, state);
- else if (has_io)
+ else if (has_io) {
+ uint64_t io_weight = cgroup_context_io_weight(c, state);
+
weight = cgroup_weight_io_to_blkio(cgroup_context_io_weight(c, state));
- else
+
+ log_cgroup_compat(u, "Applying [Startup]IOWeight %" PRIu64 " as [Startup]BlockIOWeight %" PRIu64,
+ io_weight, weight);
+ } else
weight = CGROUP_BLKIO_WEIGHT_DEFAULT;
xsprintf(buf, "%" PRIu64 "\n", weight);
r = cg_set_attribute("blkio", path, "blkio.weight", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set blkio.weight on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set blkio.weight: %m");
if (has_blockio) {
CGroupBlockIODeviceWeight *w;
/* FIXME: no way to reset this list */
LIST_FOREACH(device_weights, w, c->blockio_device_weights)
- cgroup_apply_blkio_device_weight(path, w->path, w->weight);
+ cgroup_apply_blkio_device_weight(u, w->path, w->weight);
} else if (has_io) {
CGroupIODeviceWeight *w;
/* FIXME: no way to reset this list */
- LIST_FOREACH(device_weights, w, c->io_device_weights)
- cgroup_apply_blkio_device_weight(path, w->path, cgroup_weight_io_to_blkio(w->weight));
+ LIST_FOREACH(device_weights, w, c->io_device_weights) {
+ weight = cgroup_weight_io_to_blkio(w->weight);
+
+ log_cgroup_compat(u, "Applying IODeviceWeight %" PRIu64 " as BlockIODeviceWeight %" PRIu64 " for %s",
+ w->weight, weight, w->path);
+
+ cgroup_apply_blkio_device_weight(u, w->path, weight);
+ }
}
}
@@ -674,14 +722,17 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
CGroupBlockIODeviceBandwidth *b, *next;
LIST_FOREACH_SAFE(device_bandwidths, b, next, c->blockio_device_bandwidths) {
- if (!cgroup_apply_blkio_device_limit(path, b->path, b->rbps, b->wbps))
+ if (!cgroup_apply_blkio_device_limit(u, b->path, b->rbps, b->wbps))
cgroup_context_free_blockio_device_bandwidth(c, b);
}
} else if (has_io) {
CGroupIODeviceLimit *l, *next;
LIST_FOREACH_SAFE(device_limits, l, next, c->io_device_limits) {
- if (!cgroup_apply_blkio_device_limit(path, l->path, l->limits[CGROUP_IO_RBPS_MAX], l->limits[CGROUP_IO_WBPS_MAX]))
+ log_cgroup_compat(u, "Applying IO{Read|Write}Bandwidth %" PRIu64 " %" PRIu64 " as BlockIO{Read|Write}BandwidthMax for %s",
+ l->limits[CGROUP_IO_RBPS_MAX], l->limits[CGROUP_IO_WBPS_MAX], l->path);
+
+ if (!cgroup_apply_blkio_device_limit(u, l->path, l->limits[CGROUP_IO_RBPS_MAX], l->limits[CGROUP_IO_WBPS_MAX]))
cgroup_context_free_io_device_limit(c, l);
}
}
@@ -693,24 +744,32 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
if (cgroup_context_has_unified_memory_config(c))
max = c->memory_max;
- else
+ else {
max = c->memory_limit;
- cgroup_apply_unified_memory_limit(path, "memory.low", c->memory_low);
- cgroup_apply_unified_memory_limit(path, "memory.high", c->memory_high);
- cgroup_apply_unified_memory_limit(path, "memory.max", max);
+ if (max != CGROUP_LIMIT_MAX)
+ log_cgroup_compat(u, "Applying MemoryLimit %" PRIu64 " as MemoryMax", max);
+ }
+
+ cgroup_apply_unified_memory_limit(u, "memory.low", c->memory_low);
+ cgroup_apply_unified_memory_limit(u, "memory.high", c->memory_high);
+ cgroup_apply_unified_memory_limit(u, "memory.max", max);
} else {
char buf[DECIMAL_STR_MAX(uint64_t) + 1];
if (c->memory_limit != CGROUP_LIMIT_MAX)
xsprintf(buf, "%" PRIu64 "\n", c->memory_limit);
- else
+ else {
xsprintf(buf, "%" PRIu64 "\n", c->memory_max);
+ if (c->memory_max != CGROUP_LIMIT_MAX)
+ log_cgroup_compat(u, "Applying MemoryMax %" PRIu64 " as MemoryLimit", c->memory_max);
+ }
+
r = cg_set_attribute("memory", path, "memory.limit_in_bytes", buf);
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set memory.limit_in_bytes on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set memory.limit_in_bytes: %m");
}
}
@@ -726,8 +785,8 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
else
r = cg_set_attribute("devices", path, "devices.allow", "a");
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EINVAL, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to reset devices.list on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EINVAL, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to reset devices.list: %m");
if (c->device_policy == CGROUP_CLOSED ||
(c->device_policy == CGROUP_AUTO && c->device_allow)) {
@@ -773,7 +832,7 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
else if (startswith(a->path, "char-"))
whitelist_major(path, a->path + 5, 'c', acc);
else
- log_debug("Ignoring device %s while writing cgroup attribute.", a->path);
+ log_unit_debug(u, "Ignoring device %s while writing cgroup attribute.", a->path);
}
}
@@ -788,8 +847,8 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
r = cg_set_attribute("pids", path, "pids.max", "max");
if (r < 0)
- log_full_errno(IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
- "Failed to set pids.max on %s: %m", path);
+ log_unit_full(u, IN_SET(r, -ENOENT, -EROFS, -EACCES) ? LOG_DEBUG : LOG_WARNING, r,
+ "Failed to set pids.max: %m");
}
}
@@ -1224,7 +1283,7 @@ static int unit_realize_cgroup_now(Unit *u, ManagerState state) {
return r;
/* Finally, apply the necessary attributes. */
- cgroup_context_apply(unit_get_cgroup_context(u), target_mask, u->cgroup_path, state);
+ cgroup_context_apply(u, target_mask, state);
return 0;
}
@@ -1355,7 +1414,7 @@ void unit_prune_cgroup(Unit *u) {
r = cg_trim_everywhere(u->manager->cgroup_supported, u->cgroup_path, !is_root_slice);
if (r < 0) {
- log_debug_errno(r, "Failed to destroy cgroup %s, ignoring: %m", u->cgroup_path);
+ log_unit_debug_errno(u, r, "Failed to destroy cgroup %s, ignoring: %m", u->cgroup_path);
return;
}
diff --git a/src/core/cgroup.h b/src/core/cgroup.h
index ff87adfba1..f21409bd5d 100644
--- a/src/core/cgroup.h
+++ b/src/core/cgroup.h
@@ -125,7 +125,6 @@ struct CGroupContext {
void cgroup_context_init(CGroupContext *c);
void cgroup_context_done(CGroupContext *c);
void cgroup_context_dump(CGroupContext *c, FILE* f, const char *prefix);
-void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, ManagerState state);
CGroupMask cgroup_context_get_mask(CGroupContext *c);
diff --git a/src/core/dbus-execute.c b/src/core/dbus-execute.c
index de29d5da04..4c88c41127 100644
--- a/src/core/dbus-execute.c
+++ b/src/core/dbus-execute.c
@@ -719,6 +719,7 @@ const sd_bus_vtable bus_exec_vtable[] = {
SD_BUS_PROPERTY("RestrictAddressFamilies", "(bas)", property_get_address_families, 0, SD_BUS_VTABLE_PROPERTY_CONST),
SD_BUS_PROPERTY("RuntimeDirectoryMode", "u", bus_property_get_mode, offsetof(ExecContext, runtime_directory_mode), SD_BUS_VTABLE_PROPERTY_CONST),
SD_BUS_PROPERTY("RuntimeDirectory", "as", NULL, offsetof(ExecContext, runtime_directory), SD_BUS_VTABLE_PROPERTY_CONST),
+ SD_BUS_PROPERTY("MemoryDenyWriteExecute", "b", bus_property_get_bool, offsetof(ExecContext, memory_deny_write_execute), SD_BUS_VTABLE_PROPERTY_CONST),
SD_BUS_VTABLE_END
};
@@ -1056,7 +1057,7 @@ int bus_exec_context_set_transient_property(
} else if (STR_IN_SET(name,
"IgnoreSIGPIPE", "TTYVHangup", "TTYReset",
"PrivateTmp", "PrivateDevices", "PrivateNetwork",
- "NoNewPrivileges", "SyslogLevelPrefix")) {
+ "NoNewPrivileges", "SyslogLevelPrefix", "MemoryDenyWriteExecute")) {
int b;
r = sd_bus_message_read(message, "b", &b);
@@ -1080,6 +1081,8 @@ int bus_exec_context_set_transient_property(
c->no_new_privileges = b;
else if (streq(name, "SyslogLevelPrefix"))
c->syslog_level_prefix = b;
+ else if (streq(name, "MemoryDenyWriteExecute"))
+ c->memory_deny_write_execute = b;
unit_write_drop_in_private_format(u, mode, name, "%s=%s", name, yes_no(b));
}
diff --git a/src/core/execute.c b/src/core/execute.c
index 5eb3f13695..2cef70e668 100644
--- a/src/core/execute.c
+++ b/src/core/execute.c
@@ -25,6 +25,7 @@
#include <signal.h>
#include <string.h>
#include <sys/capability.h>
+#include <sys/mman.h>
#include <sys/personality.h>
#include <sys/prctl.h>
#include <sys/socket.h>
@@ -1190,6 +1191,45 @@ finish:
return r;
}
+static int apply_memory_deny_write_execute(const ExecContext *c) {
+ scmp_filter_ctx *seccomp;
+ int r;
+
+ assert(c);
+
+ seccomp = seccomp_init(SCMP_ACT_ALLOW);
+ if (!seccomp)
+ return -ENOMEM;
+
+ r = seccomp_rule_add(
+ seccomp,
+ SCMP_ACT_KILL,
+ SCMP_SYS(mmap),
+ 1,
+ SCMP_A2(SCMP_CMP_MASKED_EQ, PROT_EXEC|PROT_WRITE, PROT_EXEC|PROT_WRITE));
+ if (r < 0)
+ goto finish;
+
+ r = seccomp_rule_add(
+ seccomp,
+ SCMP_ACT_KILL,
+ SCMP_SYS(mprotect),
+ 1,
+ SCMP_A2(SCMP_CMP_MASKED_EQ, PROT_EXEC, PROT_EXEC));
+ if (r < 0)
+ goto finish;
+
+ r = seccomp_attr_set(seccomp, SCMP_FLTATR_CTL_NNP, 0);
+ if (r < 0)
+ goto finish;
+
+ r = seccomp_load(seccomp);
+
+finish:
+ seccomp_release(seccomp);
+ return r;
+}
+
#endif
static void do_idle_pipe_dance(int idle_pipe[4]) {
@@ -1912,6 +1952,13 @@ static int exec_child(
}
}
+ if (context->memory_deny_write_execute) {
+ r = apply_memory_deny_write_execute(context);
+ if (r < 0) {
+ *exit_status = EXIT_SECCOMP;
+ return r;
+ }
+ }
if (use_syscall_filter) {
r = apply_seccomp(context);
if (r < 0) {
@@ -2371,7 +2418,8 @@ void exec_context_dump(ExecContext *c, FILE* f, const char *prefix) {
"%sPrivateDevices: %s\n"
"%sProtectHome: %s\n"
"%sProtectSystem: %s\n"
- "%sIgnoreSIGPIPE: %s\n",
+ "%sIgnoreSIGPIPE: %s\n"
+ "%sMemoryDenyWriteExecute: %s\n",
prefix, c->umask,
prefix, c->working_directory ? c->working_directory : "/",
prefix, c->root_directory ? c->root_directory : "/",
@@ -2381,7 +2429,8 @@ void exec_context_dump(ExecContext *c, FILE* f, const char *prefix) {
prefix, yes_no(c->private_devices),
prefix, protect_home_to_string(c->protect_home),
prefix, protect_system_to_string(c->protect_system),
- prefix, yes_no(c->ignore_sigpipe));
+ prefix, yes_no(c->ignore_sigpipe),
+ prefix, yes_no(c->memory_deny_write_execute));
STRV_FOREACH(e, c->environment)
fprintf(f, "%sEnvironment: %s\n", prefix, *e);
diff --git a/src/core/execute.h b/src/core/execute.h
index 41148bcea2..464869d226 100644
--- a/src/core/execute.h
+++ b/src/core/execute.h
@@ -197,6 +197,7 @@ struct ExecContext {
bool ioprio_set:1;
bool cpu_sched_set:1;
bool no_new_privileges_set:1;
+ bool memory_deny_write_execute;
};
#include "cgroup-util.h"
diff --git a/src/core/load-fragment-gperf.gperf.m4 b/src/core/load-fragment-gperf.gperf.m4
index 00bdc238ce..eb58586523 100644
--- a/src/core/load-fragment-gperf.gperf.m4
+++ b/src/core/load-fragment-gperf.gperf.m4
@@ -55,10 +55,12 @@ m4_ifdef(`HAVE_SECCOMP',
`$1.SystemCallFilter, config_parse_syscall_filter, 0, offsetof($1, exec_context)
$1.SystemCallArchitectures, config_parse_syscall_archs, 0, offsetof($1, exec_context.syscall_archs)
$1.SystemCallErrorNumber, config_parse_syscall_errno, 0, offsetof($1, exec_context)
+$1.MemoryDenyWriteExecute, config_parse_bool, 0, offsetof($1, exec_context.memory_deny_write_execute)
$1.RestrictAddressFamilies, config_parse_address_families, 0, offsetof($1, exec_context)',
`$1.SystemCallFilter, config_parse_warn_compat, DISABLED_CONFIGURATION, 0
$1.SystemCallArchitectures, config_parse_warn_compat, DISABLED_CONFIGURATION, 0
$1.SystemCallErrorNumber, config_parse_warn_compat, DISABLED_CONFIGURATION, 0
+$1.MemoryDenyWriteExecute, config_parse_warn_compat, DISABLED_CONFIGURATION, 0
$1.RestrictAddressFamilies, config_parse_warn_compat, DISABLED_CONFIGURATION, 0')
$1.LimitCPU, config_parse_limit, RLIMIT_CPU, offsetof($1, exec_context.rlimit)
$1.LimitFSIZE, config_parse_limit, RLIMIT_FSIZE, offsetof($1, exec_context.rlimit)
diff --git a/src/core/load-fragment.c b/src/core/load-fragment.c
index 09d3f65c77..b53301a147 100644
--- a/src/core/load-fragment.c
+++ b/src/core/load-fragment.c
@@ -2396,6 +2396,55 @@ int config_parse_documentation(const char *unit,
}
#ifdef HAVE_SECCOMP
+static int syscall_filter_parse_one(
+ const char *unit,
+ const char *filename,
+ unsigned line,
+ ExecContext *c,
+ bool invert,
+ const char *t,
+ bool warn) {
+ int r;
+
+ if (*t == '@') {
+ const SystemCallFilterSet *set;
+
+ for (set = syscall_filter_sets; set->set_name; set++)
+ if (streq(set->set_name, t)) {
+ const char *sys;
+
+ NULSTR_FOREACH(sys, set->value) {
+ r = syscall_filter_parse_one(unit, filename, line, c, invert, sys, false);
+ if (r < 0)
+ return r;
+ }
+ break;
+ }
+ } else {
+ int id;
+
+ id = seccomp_syscall_resolve_name(t);
+ if (id < 0) {
+ if (warn)
+ log_syntax(unit, LOG_ERR, filename, line, 0, "Failed to parse system call, ignoring: %s", t);
+ return 0;
+ }
+
+ /* If we previously wanted to forbid a syscall and now
+ * we want to allow it, then remove it from the list
+ */
+ if (!invert == c->syscall_whitelist) {
+ r = set_put(c->syscall_filter, INT_TO_PTR(id + 1));
+ if (r == 0)
+ return 0;
+ if (r < 0)
+ return log_oom();
+ } else
+ set_remove(c->syscall_filter, INT_TO_PTR(id + 1));
+ }
+ return 0;
+}
+
int config_parse_syscall_filter(
const char *unit,
const char *filename,
@@ -2408,13 +2457,6 @@ int config_parse_syscall_filter(
void *data,
void *userdata) {
- static const char default_syscalls[] =
- "execve\0"
- "exit\0"
- "exit_group\0"
- "rt_sigreturn\0"
- "sigreturn\0";
-
ExecContext *c = data;
Unit *u = userdata;
bool invert = false;
@@ -2448,53 +2490,26 @@ int config_parse_syscall_filter(
/* Allow everything but the ones listed */
c->syscall_whitelist = false;
else {
- const char *i;
-
/* Allow nothing but the ones listed */
c->syscall_whitelist = true;
/* Accept default syscalls if we are on a whitelist */
- NULSTR_FOREACH(i, default_syscalls) {
- int id;
-
- id = seccomp_syscall_resolve_name(i);
- if (id < 0)
- continue;
-
- r = set_put(c->syscall_filter, INT_TO_PTR(id + 1));
- if (r == 0)
- continue;
- if (r < 0)
- return log_oom();
- }
+ r = syscall_filter_parse_one(unit, filename, line, c, false, "@default", false);
+ if (r < 0)
+ return r;
}
}
FOREACH_WORD_QUOTED(word, l, rvalue, state) {
_cleanup_free_ char *t = NULL;
- int id;
t = strndup(word, l);
if (!t)
return log_oom();
- id = seccomp_syscall_resolve_name(t);
- if (id < 0) {
- log_syntax(unit, LOG_ERR, filename, line, 0, "Failed to parse system call, ignoring: %s", t);
- continue;
- }
-
- /* If we previously wanted to forbid a syscall and now
- * we want to allow it, then remove it from the list
- */
- if (!invert == c->syscall_whitelist) {
- r = set_put(c->syscall_filter, INT_TO_PTR(id + 1));
- if (r == 0)
- continue;
- if (r < 0)
- return log_oom();
- } else
- set_remove(c->syscall_filter, INT_TO_PTR(id + 1));
+ r = syscall_filter_parse_one(unit, filename, line, c, invert, t, true);
+ if (r < 0)
+ return r;
}
if (!isempty(state))
log_syntax(unit, LOG_ERR, filename, line, 0, "Trailing garbage, ignoring.");
@@ -2796,7 +2811,7 @@ int config_parse_memory_limit(
uint64_t bytes = CGROUP_LIMIT_MAX;
int r;
- if (!isempty(rvalue) && !streq(rvalue, "infinity") && !streq(rvalue, "max")) {
+ if (!isempty(rvalue) && !streq(rvalue, "infinity")) {
r = parse_size(rvalue, 1024, &bytes);
if (r < 0 || bytes < 1) {
log_syntax(unit, LOG_ERR, filename, line, r, "Memory limit '%s' invalid. Ignoring.", rvalue);
@@ -3065,7 +3080,7 @@ int config_parse_io_limit(
return 0;
}
- if (streq("max", limit)) {
+ if (streq("infinity", limit)) {
num = CGROUP_LIMIT_MAX;
} else {
r = parse_size(limit, 1000, &num);
generated by cgit v1.2.3-54-g00ecf (git 2.45.2) at 2025-08-17 11:10:31 +0000