summaryrefslogtreecommitdiff
path: root/src/core
diff options
context:
space:
mode:
Diffstat (limited to 'src/core')
-rw-r--r--src/core/cgroup.c360
-rw-r--r--src/core/cgroup.h30
-rw-r--r--src/core/dbus-cgroup.c331
-rw-r--r--src/core/dbus-execute.c56
-rw-r--r--src/core/dbus-job.c2
-rw-r--r--src/core/job.c24
-rw-r--r--src/core/job.h2
-rw-r--r--src/core/load-fragment-gperf.gperf.m48
-rw-r--r--src/core/load-fragment.c232
-rw-r--r--src/core/load-fragment.h3
-rw-r--r--src/core/main.c3
-rw-r--r--src/core/manager.c2
-rw-r--r--src/core/manager.h1
-rw-r--r--src/core/namespace.c12
-rw-r--r--src/core/system.conf1
-rw-r--r--src/core/transaction.c2
-rw-r--r--src/core/unit.c18
17 files changed, 955 insertions, 132 deletions
diff --git a/src/core/cgroup.c b/src/core/cgroup.c
index 996efde22b..0fb63b1bd1 100644
--- a/src/core/cgroup.c
+++ b/src/core/cgroup.c
@@ -32,6 +32,7 @@
#include "special.h"
#include "string-table.h"
#include "string-util.h"
+#include "stdio-util.h"
#define CGROUP_CPU_QUOTA_PERIOD_USEC ((usec_t) 100 * USEC_PER_MSEC)
@@ -47,6 +48,9 @@ void cgroup_context_init(CGroupContext *c) {
c->memory_limit = (uint64_t) -1;
+ c->io_weight = CGROUP_WEIGHT_INVALID;
+ c->startup_io_weight = CGROUP_WEIGHT_INVALID;
+
c->blockio_weight = CGROUP_BLKIO_WEIGHT_INVALID;
c->startup_blockio_weight = CGROUP_BLKIO_WEIGHT_INVALID;
@@ -62,6 +66,24 @@ void cgroup_context_free_device_allow(CGroupContext *c, CGroupDeviceAllow *a) {
free(a);
}
+void cgroup_context_free_io_device_weight(CGroupContext *c, CGroupIODeviceWeight *w) {
+ assert(c);
+ assert(w);
+
+ LIST_REMOVE(device_weights, c->io_device_weights, w);
+ free(w->path);
+ free(w);
+}
+
+void cgroup_context_free_io_device_limit(CGroupContext *c, CGroupIODeviceLimit *l) {
+ assert(c);
+ assert(l);
+
+ LIST_REMOVE(device_limits, c->io_device_limits, l);
+ free(l->path);
+ free(l);
+}
+
void cgroup_context_free_blockio_device_weight(CGroupContext *c, CGroupBlockIODeviceWeight *w) {
assert(c);
assert(w);
@@ -83,6 +105,12 @@ void cgroup_context_free_blockio_device_bandwidth(CGroupContext *c, CGroupBlockI
void cgroup_context_done(CGroupContext *c) {
assert(c);
+ while (c->io_device_weights)
+ cgroup_context_free_io_device_weight(c, c->io_device_weights);
+
+ while (c->io_device_limits)
+ cgroup_context_free_io_device_limit(c, c->io_device_limits);
+
while (c->blockio_device_weights)
cgroup_context_free_blockio_device_weight(c, c->blockio_device_weights);
@@ -94,6 +122,8 @@ void cgroup_context_done(CGroupContext *c) {
}
void cgroup_context_dump(CGroupContext *c, FILE* f, const char *prefix) {
+ CGroupIODeviceLimit *il;
+ CGroupIODeviceWeight *iw;
CGroupBlockIODeviceBandwidth *b;
CGroupBlockIODeviceWeight *w;
CGroupDeviceAllow *a;
@@ -106,12 +136,15 @@ void cgroup_context_dump(CGroupContext *c, FILE* f, const char *prefix) {
fprintf(f,
"%sCPUAccounting=%s\n"
+ "%sIOAccounting=%s\n"
"%sBlockIOAccounting=%s\n"
"%sMemoryAccounting=%s\n"
"%sTasksAccounting=%s\n"
"%sCPUShares=%" PRIu64 "\n"
"%sStartupCPUShares=%" PRIu64 "\n"
"%sCPUQuotaPerSecSec=%s\n"
+ "%sIOWeight=%" PRIu64 "\n"
+ "%sStartupIOWeight=%" PRIu64 "\n"
"%sBlockIOWeight=%" PRIu64 "\n"
"%sStartupBlockIOWeight=%" PRIu64 "\n"
"%sMemoryLimit=%" PRIu64 "\n"
@@ -119,12 +152,15 @@ void cgroup_context_dump(CGroupContext *c, FILE* f, const char *prefix) {
"%sDevicePolicy=%s\n"
"%sDelegate=%s\n",
prefix, yes_no(c->cpu_accounting),
+ prefix, yes_no(c->io_accounting),
prefix, yes_no(c->blockio_accounting),
prefix, yes_no(c->memory_accounting),
prefix, yes_no(c->tasks_accounting),
prefix, c->cpu_shares,
prefix, c->startup_cpu_shares,
prefix, format_timespan(u, sizeof(u), c->cpu_quota_per_sec_usec, 1),
+ prefix, c->io_weight,
+ prefix, c->startup_io_weight,
prefix, c->blockio_weight,
prefix, c->startup_blockio_weight,
prefix, c->memory_limit,
@@ -139,6 +175,27 @@ void cgroup_context_dump(CGroupContext *c, FILE* f, const char *prefix) {
a->path,
a->r ? "r" : "", a->w ? "w" : "", a->m ? "m" : "");
+ LIST_FOREACH(device_weights, iw, c->io_device_weights)
+ fprintf(f,
+ "%sIODeviceWeight=%s %" PRIu64,
+ prefix,
+ iw->path,
+ iw->weight);
+
+ LIST_FOREACH(device_limits, il, c->io_device_limits) {
+ char buf[FORMAT_BYTES_MAX];
+ CGroupIOLimitType type;
+
+ for (type = 0; type < _CGROUP_IO_LIMIT_TYPE_MAX; type++)
+ if (il->limits[type] != cgroup_io_limit_defaults[type])
+ fprintf(f,
+ "%s%s=%s %s\n",
+ prefix,
+ cgroup_io_limit_type_to_string(type),
+ il->path,
+ format_bytes(buf, sizeof(buf), il->limits[type]));
+ }
+
LIST_FOREACH(device_weights, w, c->blockio_device_weights)
fprintf(f,
"%sBlockIODeviceWeight=%s %" PRIu64,
@@ -149,16 +206,22 @@ void cgroup_context_dump(CGroupContext *c, FILE* f, const char *prefix) {
LIST_FOREACH(device_bandwidths, b, c->blockio_device_bandwidths) {
char buf[FORMAT_BYTES_MAX];
- fprintf(f,
- "%s%s=%s %s\n",
- prefix,
- b->read ? "BlockIOReadBandwidth" : "BlockIOWriteBandwidth",
- b->path,
- format_bytes(buf, sizeof(buf), b->bandwidth));
+ if (b->rbps != CGROUP_LIMIT_MAX)
+ fprintf(f,
+ "%sBlockIOReadBandwidth=%s %s\n",
+ prefix,
+ b->path,
+ format_bytes(buf, sizeof(buf), b->rbps));
+ if (b->wbps != CGROUP_LIMIT_MAX)
+ fprintf(f,
+ "%sBlockIOWriteBandwidth=%s %s\n",
+ prefix,
+ b->path,
+ format_bytes(buf, sizeof(buf), b->wbps));
}
}
-static int lookup_blkio_device(const char *p, dev_t *dev) {
+static int lookup_block_device(const char *p, dev_t *dev) {
struct stat st;
int r;
@@ -295,6 +358,144 @@ fail:
return -errno;
}
+static bool cgroup_context_has_io_config(CGroupContext *c) {
+ return c->io_accounting ||
+ c->io_weight != CGROUP_WEIGHT_INVALID ||
+ c->startup_io_weight != CGROUP_WEIGHT_INVALID ||
+ c->io_device_weights ||
+ c->io_device_limits;
+}
+
+static bool cgroup_context_has_blockio_config(CGroupContext *c) {
+ return c->blockio_accounting ||
+ c->blockio_weight != CGROUP_BLKIO_WEIGHT_INVALID ||
+ c->startup_blockio_weight != CGROUP_BLKIO_WEIGHT_INVALID ||
+ c->blockio_device_weights ||
+ c->blockio_device_bandwidths;
+}
+
+static uint64_t cgroup_context_io_weight(CGroupContext *c, ManagerState state) {
+ if (IN_SET(state, MANAGER_STARTING, MANAGER_INITIALIZING) &&
+ c->startup_io_weight != CGROUP_WEIGHT_INVALID)
+ return c->startup_io_weight;
+ else if (c->io_weight != CGROUP_WEIGHT_INVALID)
+ return c->io_weight;
+ else
+ return CGROUP_WEIGHT_DEFAULT;
+}
+
+static uint64_t cgroup_context_blkio_weight(CGroupContext *c, ManagerState state) {
+ if (IN_SET(state, MANAGER_STARTING, MANAGER_INITIALIZING) &&
+ c->startup_blockio_weight != CGROUP_BLKIO_WEIGHT_INVALID)
+ return c->startup_blockio_weight;
+ else if (c->blockio_weight != CGROUP_BLKIO_WEIGHT_INVALID)
+ return c->blockio_weight;
+ else
+ return CGROUP_BLKIO_WEIGHT_DEFAULT;
+}
+
+static uint64_t cgroup_weight_blkio_to_io(uint64_t blkio_weight) {
+ return CLAMP(blkio_weight * CGROUP_WEIGHT_DEFAULT / CGROUP_BLKIO_WEIGHT_DEFAULT,
+ CGROUP_WEIGHT_MIN, CGROUP_WEIGHT_MAX);
+}
+
+static uint64_t cgroup_weight_io_to_blkio(uint64_t io_weight) {
+ return CLAMP(io_weight * CGROUP_BLKIO_WEIGHT_DEFAULT / CGROUP_WEIGHT_DEFAULT,
+ 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) {
+ char buf[DECIMAL_STR_MAX(dev_t)*2+2+DECIMAL_STR_MAX(uint64_t)+1];
+ dev_t dev;
+ int r;
+
+ r = lookup_block_device(dev_path, &dev);
+ if (r < 0)
+ return;
+
+ xsprintf(buf, "%u:%u %" PRIu64 "\n", major(dev), minor(dev), io_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);
+}
+
+static void cgroup_apply_blkio_device_weight(const char *path, 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;
+
+ r = lookup_block_device(dev_path, &dev);
+ if (r < 0)
+ return;
+
+ xsprintf(buf, "%u:%u %" PRIu64 "\n", major(dev), minor(dev), blkio_weight);
+ r = cg_set_attribute("blkio", 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);
+}
+
+static unsigned cgroup_apply_io_device_limit(const char *path, 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;
+ dev_t dev;
+ unsigned n = 0;
+ int r;
+
+ r = lookup_block_device(dev_path, &dev);
+ if (r < 0)
+ return 0;
+
+ for (type = 0; type < _CGROUP_IO_LIMIT_TYPE_MAX; type++) {
+ if (limits[type] != cgroup_io_limit_defaults[type]) {
+ xsprintf(limit_bufs[type], "%" PRIu64, limits[type]);
+ n++;
+ } else {
+ xsprintf(limit_bufs[type], "%s", limits[type] == CGROUP_LIMIT_MAX ? "max" : "0");
+ }
+ }
+
+ 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);
+ 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);
+ return n;
+}
+
+static unsigned cgroup_apply_blkio_device_limit(const char *path, 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;
+ int r;
+
+ r = lookup_block_device(dev_path, &dev);
+ if (r < 0)
+ return 0;
+
+ 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);
+ 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);
+
+ 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);
+ 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);
+
+ return n;
+}
+
void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, ManagerState state) {
bool is_root;
int r;
@@ -343,53 +544,120 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
"Failed to set cpu.cfs_quota_us on %s: %m", path);
}
- if (mask & CGROUP_MASK_BLKIO) {
- char buf[MAX(DECIMAL_STR_MAX(uint64_t)+1,
- DECIMAL_STR_MAX(dev_t)*2+2+DECIMAL_STR_MAX(uint64_t)+1)];
- CGroupBlockIODeviceWeight *w;
- CGroupBlockIODeviceBandwidth *b;
+ if (mask & CGROUP_MASK_IO) {
+ bool has_io = cgroup_context_has_io_config(c);
+ bool has_blockio = cgroup_context_has_blockio_config(c);
if (!is_root) {
- sprintf(buf, "%" PRIu64 "\n",
- IN_SET(state, MANAGER_STARTING, MANAGER_INITIALIZING) && c->startup_blockio_weight != CGROUP_BLKIO_WEIGHT_INVALID ? c->startup_blockio_weight :
- c->blockio_weight != CGROUP_BLKIO_WEIGHT_INVALID ? c->blockio_weight : CGROUP_BLKIO_WEIGHT_DEFAULT);
- r = cg_set_attribute("blkio", path, "blkio.weight", buf);
+ char buf[8+DECIMAL_STR_MAX(uint64_t)+1];
+ uint64_t weight;
+
+ 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
+ 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 blkio.weight on %s: %m", path);
+ "Failed to set io.weight on %s: %m", path);
- /* FIXME: no way to reset this list */
- LIST_FOREACH(device_weights, w, c->blockio_device_weights) {
- dev_t dev;
+ if (has_io) {
+ CGroupIODeviceWeight *w;
- r = lookup_blkio_device(w->path, &dev);
- if (r < 0)
- continue;
+ /* 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);
+ } else if (has_blockio) {
+ CGroupBlockIODeviceWeight *w;
- sprintf(buf, "%u:%u %" PRIu64 "\n", major(dev), minor(dev), w->weight);
- r = cg_set_attribute("blkio", 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);
+ /* 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));
}
}
- /* FIXME: no way to reset this list */
- LIST_FOREACH(device_bandwidths, b, c->blockio_device_bandwidths) {
- const char *a;
- dev_t dev;
+ /* Apply limits and free ones without config. */
+ if (has_io) {
+ CGroupIODeviceLimit *l, *next;
- r = lookup_blkio_device(b->path, &dev);
- if (r < 0)
- continue;
+ LIST_FOREACH_SAFE(device_limits, l, next, c->io_device_limits) {
+ if (!cgroup_apply_io_device_limit(path, l->path, l->limits))
+ cgroup_context_free_io_device_limit(c, l);
+ }
+ } else if (has_blockio) {
+ CGroupBlockIODeviceBandwidth *b, *next;
+
+ LIST_FOREACH_SAFE(device_bandwidths, b, next, c->blockio_device_bandwidths) {
+ uint64_t limits[_CGROUP_IO_LIMIT_TYPE_MAX];
+ CGroupIOLimitType type;
+
+ for (type = 0; type < _CGROUP_IO_LIMIT_TYPE_MAX; type++)
+ limits[type] = cgroup_io_limit_defaults[type];
+
+ limits[CGROUP_IO_RBPS_MAX] = b->rbps;
+ limits[CGROUP_IO_WBPS_MAX] = b->wbps;
+
+ if (!cgroup_apply_io_device_limit(path, b->path, limits))
+ cgroup_context_free_blockio_device_bandwidth(c, b);
+ }
+ }
+ }
+
+ if (mask & CGROUP_MASK_BLKIO) {
+ bool has_io = cgroup_context_has_io_config(c);
+ bool has_blockio = cgroup_context_has_blockio_config(c);
+
+ if (!is_root) {
+ char buf[DECIMAL_STR_MAX(uint64_t)+1];
+ uint64_t weight;
- a = b->read ? "blkio.throttle.read_bps_device" : "blkio.throttle.write_bps_device";
+ if (has_blockio)
+ weight = cgroup_context_blkio_weight(c, state);
+ else if (has_io)
+ weight = cgroup_weight_io_to_blkio(cgroup_context_io_weight(c, state));
+ else
+ weight = CGROUP_BLKIO_WEIGHT_DEFAULT;
- sprintf(buf, "%u:%u %" PRIu64 "\n", major(dev), minor(dev), b->bandwidth);
- r = cg_set_attribute("blkio", path, a, buf);
+ 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 %s on %s: %m", a, path);
+ "Failed to set blkio.weight on %s: %m", path);
+
+ 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);
+ } 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));
+ }
+ }
+
+ /* Apply limits and free ones without config. */
+ if (has_blockio) {
+ 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))
+ 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]))
+ cgroup_context_free_io_device_limit(c, l);
+ }
}
}
@@ -506,12 +774,8 @@ CGroupMask cgroup_context_get_mask(CGroupContext *c) {
c->cpu_quota_per_sec_usec != USEC_INFINITY)
mask |= CGROUP_MASK_CPUACCT | CGROUP_MASK_CPU;
- if (c->blockio_accounting ||
- c->blockio_weight != CGROUP_BLKIO_WEIGHT_INVALID ||
- c->startup_blockio_weight != CGROUP_BLKIO_WEIGHT_INVALID ||
- c->blockio_device_weights ||
- c->blockio_device_bandwidths)
- mask |= CGROUP_MASK_BLKIO;
+ if (cgroup_context_has_io_config(c) || cgroup_context_has_blockio_config(c))
+ mask |= CGROUP_MASK_IO | CGROUP_MASK_BLKIO;
if (c->memory_accounting ||
c->memory_limit != (uint64_t) -1)
@@ -1601,6 +1865,10 @@ void unit_invalidate_cgroup(Unit *u, CGroupMask m) {
if (m == 0)
return;
+ /* always invalidate compat pairs together */
+ if (m & (CGROUP_MASK_IO | CGROUP_MASK_BLKIO))
+ m |= CGROUP_MASK_IO | CGROUP_MASK_BLKIO;
+
if ((u->cgroup_realized_mask & m) == 0)
return;
@@ -1615,7 +1883,7 @@ void manager_invalidate_startup_units(Manager *m) {
assert(m);
SET_FOREACH(u, m->startup_units, i)
- unit_invalidate_cgroup(u, CGROUP_MASK_CPU|CGROUP_MASK_BLKIO);
+ unit_invalidate_cgroup(u, CGROUP_MASK_CPU|CGROUP_MASK_IO|CGROUP_MASK_BLKIO);
}
static const char* const cgroup_device_policy_table[_CGROUP_DEVICE_POLICY_MAX] = {
diff --git a/src/core/cgroup.h b/src/core/cgroup.h
index 360bbca30f..2b1edbafc4 100644
--- a/src/core/cgroup.h
+++ b/src/core/cgroup.h
@@ -23,9 +23,12 @@
#include "list.h"
#include "time-util.h"
+#include "cgroup-util.h"
typedef struct CGroupContext CGroupContext;
typedef struct CGroupDeviceAllow CGroupDeviceAllow;
+typedef struct CGroupIODeviceWeight CGroupIODeviceWeight;
+typedef struct CGroupIODeviceLimit CGroupIODeviceLimit;
typedef struct CGroupBlockIODeviceWeight CGroupBlockIODeviceWeight;
typedef struct CGroupBlockIODeviceBandwidth CGroupBlockIODeviceBandwidth;
@@ -53,6 +56,18 @@ struct CGroupDeviceAllow {
bool m:1;
};
+struct CGroupIODeviceWeight {
+ LIST_FIELDS(CGroupIODeviceWeight, device_weights);
+ char *path;
+ uint64_t weight;
+};
+
+struct CGroupIODeviceLimit {
+ LIST_FIELDS(CGroupIODeviceLimit, device_limits);
+ char *path;
+ uint64_t limits[_CGROUP_IO_LIMIT_TYPE_MAX];
+};
+
struct CGroupBlockIODeviceWeight {
LIST_FIELDS(CGroupBlockIODeviceWeight, device_weights);
char *path;
@@ -62,16 +77,24 @@ struct CGroupBlockIODeviceWeight {
struct CGroupBlockIODeviceBandwidth {
LIST_FIELDS(CGroupBlockIODeviceBandwidth, device_bandwidths);
char *path;
- uint64_t bandwidth;
- bool read;
+ uint64_t rbps;
+ uint64_t wbps;
};
struct CGroupContext {
bool cpu_accounting;
+ bool io_accounting;
bool blockio_accounting;
bool memory_accounting;
bool tasks_accounting;
+ /* For unified hierarchy */
+ uint64_t io_weight;
+ uint64_t startup_io_weight;
+ LIST_HEAD(CGroupIODeviceWeight, io_device_weights);
+ LIST_HEAD(CGroupIODeviceLimit, io_device_limits);
+
+ /* For legacy hierarchies */
uint64_t cpu_shares;
uint64_t startup_cpu_shares;
usec_t cpu_quota_per_sec_usec;
@@ -86,6 +109,7 @@ struct CGroupContext {
CGroupDevicePolicy device_policy;
LIST_HEAD(CGroupDeviceAllow, device_allow);
+ /* Common */
uint64_t tasks_max;
bool delegate;
@@ -102,6 +126,8 @@ void cgroup_context_apply(CGroupContext *c, CGroupMask mask, const char *path, M
CGroupMask cgroup_context_get_mask(CGroupContext *c);
void cgroup_context_free_device_allow(CGroupContext *c, CGroupDeviceAllow *a);
+void cgroup_context_free_io_device_weight(CGroupContext *c, CGroupIODeviceWeight *w);
+void cgroup_context_free_io_device_limit(CGroupContext *c, CGroupIODeviceLimit *l);
void cgroup_context_free_blockio_device_weight(CGroupContext *c, CGroupBlockIODeviceWeight *w);
void cgroup_context_free_blockio_device_bandwidth(CGroupContext *c, CGroupBlockIODeviceBandwidth *b);
diff --git a/src/core/dbus-cgroup.c b/src/core/dbus-cgroup.c
index 859d155ec1..eef1c47c14 100644
--- a/src/core/dbus-cgroup.c
+++ b/src/core/dbus-cgroup.c
@@ -28,6 +28,72 @@
static BUS_DEFINE_PROPERTY_GET_ENUM(property_get_cgroup_device_policy, cgroup_device_policy, CGroupDevicePolicy);
+static int property_get_io_device_weight(
+ sd_bus *bus,
+ const char *path,
+ const char *interface,
+ const char *property,
+ sd_bus_message *reply,
+ void *userdata,
+ sd_bus_error *error) {
+
+ CGroupContext *c = userdata;
+ CGroupIODeviceWeight *w;
+ int r;
+
+ assert(bus);
+ assert(reply);
+ assert(c);
+
+ r = sd_bus_message_open_container(reply, 'a', "(st)");
+ if (r < 0)
+ return r;
+
+ LIST_FOREACH(device_weights, w, c->io_device_weights) {
+ r = sd_bus_message_append(reply, "(st)", w->path, w->weight);
+ if (r < 0)
+ return r;
+ }
+
+ return sd_bus_message_close_container(reply);
+}
+
+static int property_get_io_device_limits(
+ sd_bus *bus,
+ const char *path,
+ const char *interface,
+ const char *property,
+ sd_bus_message *reply,
+ void *userdata,
+ sd_bus_error *error) {
+
+ CGroupContext *c = userdata;
+ CGroupIODeviceLimit *l;
+ int r;
+
+ assert(bus);
+ assert(reply);
+ assert(c);
+
+ r = sd_bus_message_open_container(reply, 'a', "(st)");
+ if (r < 0)
+ return r;
+
+ LIST_FOREACH(device_limits, l, c->io_device_limits) {
+ CGroupIOLimitType type;
+
+ type = cgroup_io_limit_type_from_string(property);
+ if (type < 0 || l->limits[type] == cgroup_io_limit_defaults[type])
+ continue;
+
+ r = sd_bus_message_append(reply, "(st)", l->path, l->limits[type]);
+ if (r < 0)
+ return r;
+ }
+
+ return sd_bus_message_close_container(reply);
+}
+
static int property_get_blockio_device_weight(
sd_bus *bus,
const char *path,
@@ -80,11 +146,17 @@ static int property_get_blockio_device_bandwidths(
return r;
LIST_FOREACH(device_bandwidths, b, c->blockio_device_bandwidths) {
+ uint64_t v;
- if (streq(property, "BlockIOReadBandwidth") != b->read)
+ if (streq(property, "BlockIOReadBandwidth"))
+ v = b->rbps;
+ else
+ v = b->wbps;
+
+ if (v == CGROUP_LIMIT_MAX)
continue;
- r = sd_bus_message_append(reply, "(st)", b->path, b->bandwidth);
+ r = sd_bus_message_append(reply, "(st)", b->path, v);
if (r < 0)
return r;
}
@@ -141,6 +213,14 @@ const sd_bus_vtable bus_cgroup_vtable[] = {
SD_BUS_PROPERTY("CPUShares", "t", NULL, offsetof(CGroupContext, cpu_shares), 0),
SD_BUS_PROPERTY("StartupCPUShares", "t", NULL, offsetof(CGroupContext, startup_cpu_shares), 0),
SD_BUS_PROPERTY("CPUQuotaPerSecUSec", "t", bus_property_get_usec, offsetof(CGroupContext, cpu_quota_per_sec_usec), 0),
+ SD_BUS_PROPERTY("IOAccounting", "b", bus_property_get_bool, offsetof(CGroupContext, io_accounting), 0),
+ SD_BUS_PROPERTY("IOWeight", "t", NULL, offsetof(CGroupContext, io_weight), 0),
+ SD_BUS_PROPERTY("StartupIOWeight", "t", NULL, offsetof(CGroupContext, startup_io_weight), 0),
+ SD_BUS_PROPERTY("IODeviceWeight", "a(st)", property_get_io_device_weight, 0, 0),
+ SD_BUS_PROPERTY("IOReadBandwidthMax", "a(st)", property_get_io_device_limits, 0, 0),
+ SD_BUS_PROPERTY("IOWriteBandwidthMax", "a(st)", property_get_io_device_limits, 0, 0),
+ SD_BUS_PROPERTY("IOReadIOPSMax", "a(st)", property_get_io_device_limits, 0, 0),
+ SD_BUS_PROPERTY("IOWriteIOPSMax", "a(st)", property_get_io_device_limits, 0, 0),
SD_BUS_PROPERTY("BlockIOAccounting", "b", bus_property_get_bool, offsetof(CGroupContext, blockio_accounting), 0),
SD_BUS_PROPERTY("BlockIOWeight", "t", NULL, offsetof(CGroupContext, blockio_weight), 0),
SD_BUS_PROPERTY("StartupBlockIOWeight", "t", NULL, offsetof(CGroupContext, startup_blockio_weight), 0),
@@ -197,6 +277,7 @@ int bus_cgroup_set_property(
UnitSetPropertiesMode mode,
sd_bus_error *error) {
+ CGroupIOLimitType iol_type;
int r;
assert(u);
@@ -281,6 +362,223 @@ int bus_cgroup_set_property(
return 1;
+ } else if (streq(name, "IOAccounting")) {
+ int b;
+
+ r = sd_bus_message_read(message, "b", &b);
+ if (r < 0)
+ return r;
+
+ if (mode != UNIT_CHECK) {
+ c->io_accounting = b;
+ unit_invalidate_cgroup(u, CGROUP_MASK_IO);
+ unit_write_drop_in_private(u, mode, name, b ? "IOAccounting=yes" : "IOAccounting=no");
+ }
+
+ return 1;
+
+ } else if (streq(name, "IOWeight")) {
+ uint64_t weight;
+
+ r = sd_bus_message_read(message, "t", &weight);
+ if (r < 0)
+ return r;
+
+ if (!CGROUP_WEIGHT_IS_OK(weight))
+ return sd_bus_error_set_errnof(error, EINVAL, "IOWeight value out of range");
+
+ if (mode != UNIT_CHECK) {
+ c->io_weight = weight;
+ unit_invalidate_cgroup(u, CGROUP_MASK_IO);
+
+ if (weight == CGROUP_WEIGHT_INVALID)
+ unit_write_drop_in_private(u, mode, name, "IOWeight=");
+ else
+ unit_write_drop_in_private_format(u, mode, name, "IOWeight=%" PRIu64, weight);
+ }
+
+ return 1;
+
+ } else if (streq(name, "StartupIOWeight")) {
+ uint64_t weight;
+
+ r = sd_bus_message_read(message, "t", &weight);
+ if (r < 0)
+ return r;
+
+ if (CGROUP_WEIGHT_IS_OK(weight))
+ return sd_bus_error_set_errnof(error, EINVAL, "StartupIOWeight value out of range");
+
+ if (mode != UNIT_CHECK) {
+ c->startup_io_weight = weight;
+ unit_invalidate_cgroup(u, CGROUP_MASK_IO);
+
+ if (weight == CGROUP_WEIGHT_INVALID)
+ unit_write_drop_in_private(u, mode, name, "StartupIOWeight=");
+ else
+ unit_write_drop_in_private_format(u, mode, name, "StartupIOWeight=%" PRIu64, weight);
+ }
+
+ return 1;
+
+ } else if ((iol_type = cgroup_io_limit_type_from_string(name)) >= 0) {
+ const char *path;
+ unsigned n = 0;
+ uint64_t u64;
+
+ r = sd_bus_message_enter_container(message, 'a', "(st)");
+ if (r < 0)
+ return r;
+
+ while ((r = sd_bus_message_read(message, "(st)", &path, &u64)) > 0) {
+
+ if (mode != UNIT_CHECK) {
+ CGroupIODeviceLimit *a = NULL, *b;
+
+ LIST_FOREACH(device_limits, b, c->io_device_limits) {
+ if (path_equal(path, b->path)) {
+ a = b;
+ break;
+ }
+ }
+
+ if (!a) {
+ CGroupIOLimitType type;
+
+ a = new0(CGroupIODeviceLimit, 1);
+ if (!a)
+ return -ENOMEM;
+
+ a->path = strdup(path);
+ if (!a->path) {
+ free(a);
+ return -ENOMEM;
+ }
+
+ for (type = 0; type < _CGROUP_IO_LIMIT_TYPE_MAX; type++)
+ a->limits[type] = cgroup_io_limit_defaults[type];
+
+ LIST_PREPEND(device_limits, c->io_device_limits, a);
+ }
+
+ a->limits[iol_type] = u64;
+ }
+
+ n++;
+ }
+ if (r < 0)
+ return r;
+
+ r = sd_bus_message_exit_container(message);
+ if (r < 0)
+ return r;
+
+ if (mode != UNIT_CHECK) {
+ CGroupIODeviceLimit *a;
+ _cleanup_free_ char *buf = NULL;
+ _cleanup_fclose_ FILE *f = NULL;
+ size_t size = 0;
+
+ if (n == 0) {
+ LIST_FOREACH(device_limits, a, c->io_device_limits)
+ a->limits[iol_type] = cgroup_io_limit_defaults[iol_type];
+ }
+
+ unit_invalidate_cgroup(u, CGROUP_MASK_IO);
+
+ f = open_memstream(&buf, &size);
+ if (!f)
+ return -ENOMEM;
+
+ fprintf(f, "%s=\n", name);
+ LIST_FOREACH(device_limits, a, c->io_device_limits)
+ if (a->limits[iol_type] != cgroup_io_limit_defaults[iol_type])
+ fprintf(f, "%s=%s %" PRIu64 "\n", name, a->path, a->limits[iol_type]);
+
+ r = fflush_and_check(f);
+ if (r < 0)
+ return r;
+ unit_write_drop_in_private(u, mode, name, buf);
+ }
+
+ return 1;
+
+ } else if (streq(name, "IODeviceWeight")) {
+ const char *path;
+ uint64_t weight;
+ unsigned n = 0;
+
+ r = sd_bus_message_enter_container(message, 'a', "(st)");
+ if (r < 0)
+ return r;
+
+ while ((r = sd_bus_message_read(message, "(st)", &path, &weight)) > 0) {
+
+ if (!CGROUP_WEIGHT_IS_OK(weight) || weight == CGROUP_WEIGHT_INVALID)
+ return sd_bus_error_set_errnof(error, EINVAL, "IODeviceWeight out of range");
+
+ if (mode != UNIT_CHECK) {
+ CGroupIODeviceWeight *a = NULL, *b;
+
+ LIST_FOREACH(device_weights, b, c->io_device_weights) {
+ if (path_equal(b->path, path)) {
+ a = b;
+ break;
+ }
+ }
+
+ if (!a) {
+ a = new0(CGroupIODeviceWeight, 1);
+ if (!a)
+ return -ENOMEM;
+
+ a->path = strdup(path);
+ if (!a->path) {
+ free(a);
+ return -ENOMEM;
+ }
+ LIST_PREPEND(device_weights,c->io_device_weights, a);
+ }
+
+ a->weight = weight;
+ }
+
+ n++;
+ }
+
+ r = sd_bus_message_exit_container(message);
+ if (r < 0)
+ return r;
+
+ if (mode != UNIT_CHECK) {
+ _cleanup_free_ char *buf = NULL;
+ _cleanup_fclose_ FILE *f = NULL;
+ CGroupIODeviceWeight *a;
+ size_t size = 0;
+
+ if (n == 0) {
+ while (c->io_device_weights)
+ cgroup_context_free_io_device_weight(c, c->io_device_weights);
+ }
+
+ unit_invalidate_cgroup(u, CGROUP_MASK_IO);
+
+ f = open_memstream(&buf, &size);
+ if (!f)
+ return -ENOMEM;
+
+ fputs("IODeviceWeight=\n", f);
+ LIST_FOREACH(device_weights, a, c->io_device_weights)
+ fprintf(f, "IODeviceWeight=%s %" PRIu64 "\n", a->path, a->weight);
+
+ r = fflush_and_check(f);
+ if (r < 0)
+ return r;
+ unit_write_drop_in_private(u, mode, name, buf);
+ }
+
+ return 1;
+
} else if (streq(name, "BlockIOAccounting")) {
int b;
@@ -359,7 +657,7 @@ int bus_cgroup_set_property(
CGroupBlockIODeviceBandwidth *a = NULL, *b;
LIST_FOREACH(device_bandwidths, b, c->blockio_device_bandwidths) {
- if (path_equal(path, b->path) && read == b->read) {
+ if (path_equal(path, b->path)) {
a = b;
break;
}
@@ -370,7 +668,8 @@ int bus_cgroup_set_property(
if (!a)
return -ENOMEM;
- a->read = read;
+ a->rbps = CGROUP_LIMIT_MAX;
+ a->wbps = CGROUP_LIMIT_MAX;
a->path = strdup(path);
if (!a->path) {
free(a);
@@ -380,7 +679,10 @@ int bus_cgroup_set_property(
LIST_PREPEND(device_bandwidths, c->blockio_device_bandwidths, a);
}
- a->bandwidth = u64;
+ if (read)
+ a->rbps = u64;
+ else
+ a->wbps = u64;
}
n++;
@@ -393,15 +695,18 @@ int bus_cgroup_set_property(
return r;
if (mode != UNIT_CHECK) {
- CGroupBlockIODeviceBandwidth *a, *next;
+ CGroupBlockIODeviceBandwidth *a;
_cleanup_free_ char *buf = NULL;
_cleanup_fclose_ FILE *f = NULL;
size_t size = 0;
if (n == 0) {
- LIST_FOREACH_SAFE(device_bandwidths, a, next, c->blockio_device_bandwidths)
- if (a->read == read)
- cgroup_context_free_blockio_device_bandwidth(c, a);
+ LIST_FOREACH(device_bandwidths, a, c->blockio_device_bandwidths) {
+ if (read)
+ a->rbps = CGROUP_LIMIT_MAX;
+ else
+ a->wbps = CGROUP_LIMIT_MAX;
+ }
}
unit_invalidate_cgroup(u, CGROUP_MASK_BLKIO);
@@ -413,13 +718,13 @@ int bus_cgroup_set_property(
if (read) {
fputs("BlockIOReadBandwidth=\n", f);
LIST_FOREACH(device_bandwidths, a, c->blockio_device_bandwidths)
- if (a->read)
- fprintf(f, "BlockIOReadBandwidth=%s %" PRIu64 "\n", a->path, a->bandwidth);
+ if (a->rbps != CGROUP_LIMIT_MAX)
+ fprintf(f, "BlockIOReadBandwidth=%s %" PRIu64 "\n", a->path, a->rbps);
} else {
fputs("BlockIOWriteBandwidth=\n", f);
LIST_FOREACH(device_bandwidths, a, c->blockio_device_bandwidths)
- if (!a->read)
- fprintf(f, "BlockIOWriteBandwidth=%s %" PRIu64 "\n", a->path, a->bandwidth);
+ if (a->wbps != CGROUP_LIMIT_MAX)
+ fprintf(f, "BlockIOWriteBandwidth=%s %" PRIu64 "\n", a->path, a->wbps);
}
r = fflush_and_check(f);
diff --git a/src/core/dbus-execute.c b/src/core/dbus-execute.c
index 9dfca14914..06943c6365 100644
--- a/src/core/dbus-execute.c
+++ b/src/core/dbus-execute.c
@@ -839,16 +839,8 @@ int bus_exec_context_set_transient_property(
if (isempty(uu))
c->user = mfree(c->user);
- else {
- char *t;
-
- t = strdup(uu);
- if (!t)
- return -ENOMEM;
-
- free(c->user);
- c->user = t;
- }
+ else if (free_and_strdup(&c->user, uu) < 0)
+ return -ENOMEM;
unit_write_drop_in_private_format(u, mode, name, "User=%s\n", uu);
}
@@ -866,16 +858,8 @@ int bus_exec_context_set_transient_property(
if (isempty(gg))
c->group = mfree(c->group);
- else {
- char *t;
-
- t = strdup(gg);
- if (!t)
- return -ENOMEM;
-
- free(c->group);
- c->group = t;
- }
+ else if (free_and_strdup(&c->group, gg) < 0)
+ return -ENOMEM;
unit_write_drop_in_private_format(u, mode, name, "Group=%s\n", gg);
}
@@ -890,18 +874,10 @@ int bus_exec_context_set_transient_property(
if (mode != UNIT_CHECK) {
- if (isempty(id)) {
+ if (isempty(id))
c->syslog_identifier = mfree(c->syslog_identifier);
- } else {
- char *t;
-
- t = strdup(id);
- if (!t)
- return -ENOMEM;
-
- free(c->syslog_identifier);
- c->syslog_identifier = t;
- }
+ else if (free_and_strdup(&c->syslog_identifier, id) < 0)
+ return -ENOMEM;
unit_write_drop_in_private_format(u, mode, name, "SyslogIdentifier=%s\n", id);
}
@@ -1478,6 +1454,24 @@ int bus_exec_context_set_transient_property(
return 1;
+ } else if (streq(name, "SELinuxContext")) {
+ const char *s;
+
+ r = sd_bus_message_read(message, "s", &s);
+ if (r < 0)
+ return r;
+
+ if (mode != UNIT_CHECK) {
+ if (isempty(s))
+ c->selinux_context = mfree(c->selinux_context);
+ else if (free_and_strdup(&c->selinux_context, s) < 0)
+ return -ENOMEM;
+
+ unit_write_drop_in_private_format(u, mode, name, "%s=%s\n", name, strempty(s));
+ }
+
+ return 1;
+
}
ri = rlimit_from_string(name);
diff --git a/src/core/dbus-job.c b/src/core/dbus-job.c
index 97a93fb2f1..ccf7453d47 100644
--- a/src/core/dbus-job.c
+++ b/src/core/dbus-job.c
@@ -75,7 +75,7 @@ int bus_job_method_cancel(sd_bus_message *message, void *userdata, sd_bus_error
return 1; /* No authorization for now, but the async polkit stuff will call us again when it has it */
}
- job_finish_and_invalidate(j, JOB_CANCELED, true);
+ job_finish_and_invalidate(j, JOB_CANCELED, true, false);
return sd_bus_reply_method_return(message, NULL);
}
diff --git a/src/core/job.c b/src/core/job.c
index d9c5669c9f..7557874d4d 100644
--- a/src/core/job.c
+++ b/src/core/job.c
@@ -191,7 +191,7 @@ Job* job_install(Job *j) {
if (uj) {
if (job_type_is_conflicting(uj->type, j->type))
- job_finish_and_invalidate(uj, JOB_CANCELED, false);
+ job_finish_and_invalidate(uj, JOB_CANCELED, false, false);
else {
/* not conflicting, i.e. mergeable */
@@ -614,19 +614,19 @@ int job_run_and_invalidate(Job *j) {
if (j) {
if (r == -EALREADY)
- r = job_finish_and_invalidate(j, JOB_DONE, true);
+ r = job_finish_and_invalidate(j, JOB_DONE, true, true);
else if (r == -EBADR)
- r = job_finish_and_invalidate(j, JOB_SKIPPED, true);
+ r = job_finish_and_invalidate(j, JOB_SKIPPED, true, false);
else if (r == -ENOEXEC)
- r = job_finish_and_invalidate(j, JOB_INVALID, true);
+ r = job_finish_and_invalidate(j, JOB_INVALID, true, false);
else if (r == -EPROTO)
- r = job_finish_and_invalidate(j, JOB_ASSERT, true);
+ r = job_finish_and_invalidate(j, JOB_ASSERT, true, false);
else if (r == -EOPNOTSUPP)
- r = job_finish_and_invalidate(j, JOB_UNSUPPORTED, true);
+ r = job_finish_and_invalidate(j, JOB_UNSUPPORTED, true, false);
else if (r == -EAGAIN)
job_set_state(j, JOB_WAITING);
else if (r < 0)
- r = job_finish_and_invalidate(j, JOB_FAILED, true);
+ r = job_finish_and_invalidate(j, JOB_FAILED, true, false);
}
return r;
@@ -827,11 +827,11 @@ static void job_fail_dependencies(Unit *u, UnitDependency d) {
if (!IN_SET(j->type, JOB_START, JOB_VERIFY_ACTIVE))
continue;
- job_finish_and_invalidate(j, JOB_DEPENDENCY, true);
+ job_finish_and_invalidate(j, JOB_DEPENDENCY, true, false);
}
}
-int job_finish_and_invalidate(Job *j, JobResult result, bool recursive) {
+int job_finish_and_invalidate(Job *j, JobResult result, bool recursive, bool already) {
Unit *u;
Unit *other;
JobType t;
@@ -848,7 +848,9 @@ int job_finish_and_invalidate(Job *j, JobResult result, bool recursive) {
log_unit_debug(u, "Job %s/%s finished, result=%s", u->id, job_type_to_string(t), job_result_to_string(result));
- job_emit_status_message(u, t, result);
+ /* If this job did nothing to respective unit we don't log the status message */
+ if (!already)
+ job_emit_status_message(u, t, result);
job_add_to_dbus_queue(j);
@@ -923,7 +925,7 @@ static int job_dispatch_timer(sd_event_source *s, uint64_t monotonic, void *user
log_unit_warning(j->unit, "Job %s/%s timed out.", j->unit->id, job_type_to_string(j->type));
u = j->unit;
- job_finish_and_invalidate(j, JOB_TIMEOUT, true);
+ job_finish_and_invalidate(j, JOB_TIMEOUT, true, false);
failure_action(u->manager, u->job_timeout_action, u->job_timeout_reboot_arg);
diff --git a/src/core/job.h b/src/core/job.h
index 856b0ce829..d359e8bb3e 100644
--- a/src/core/job.h
+++ b/src/core/job.h
@@ -219,7 +219,7 @@ void job_add_to_dbus_queue(Job *j);
int job_start_timer(Job *j);
int job_run_and_invalidate(Job *j);
-int job_finish_and_invalidate(Job *j, JobResult result, bool recursive);
+int job_finish_and_invalidate(Job *j, JobResult result, bool recursive, bool already);
char *job_dbus_path(Job *j);
diff --git a/src/core/load-fragment-gperf.gperf.m4 b/src/core/load-fragment-gperf.gperf.m4
index 928b913c7b..8193418980 100644
--- a/src/core/load-fragment-gperf.gperf.m4
+++ b/src/core/load-fragment-gperf.gperf.m4
@@ -120,6 +120,14 @@ $1.MemoryAccounting, config_parse_bool, 0,
$1.MemoryLimit, config_parse_memory_limit, 0, offsetof($1, cgroup_context)
$1.DeviceAllow, config_parse_device_allow, 0, offsetof($1, cgroup_context)
$1.DevicePolicy, config_parse_device_policy, 0, offsetof($1, cgroup_context.device_policy)
+$1.IOAccounting, config_parse_bool, 0, offsetof($1, cgroup_context.io_accounting)
+$1.IOWeight, config_parse_io_weight, 0, offsetof($1, cgroup_context.io_weight)
+$1.StartupIOWeight, config_parse_io_weight, 0, offsetof($1, cgroup_context.startup_io_weight)
+$1.IODeviceWeight, config_parse_io_device_weight, 0, offsetof($1, cgroup_context)
+$1.IOReadBandwidthMax, config_parse_io_limit, 0, offsetof($1, cgroup_context)
+$1.IOWriteBandwidthMax, config_parse_io_limit, 0, offsetof($1, cgroup_context)
+$1.IOReadIOPSMax, config_parse_io_limit, 0, offsetof($1, cgroup_context)
+$1.IOWriteIOPSMax, config_parse_io_limit, 0, offsetof($1, cgroup_context)
$1.BlockIOAccounting, config_parse_bool, 0, offsetof($1, cgroup_context.blockio_accounting)
$1.BlockIOWeight, config_parse_blockio_weight, 0, offsetof($1, cgroup_context.blockio_weight)
$1.StartupBlockIOWeight, config_parse_blockio_weight, 0, offsetof($1, cgroup_context.startup_blockio_weight)
diff --git a/src/core/load-fragment.c b/src/core/load-fragment.c
index a12dd38c60..86b4fb071b 100644
--- a/src/core/load-fragment.c
+++ b/src/core/load-fragment.c
@@ -2910,6 +2910,193 @@ int config_parse_device_allow(
return 0;
}
+int config_parse_io_weight(
+ 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) {
+
+ uint64_t *weight = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ r = cg_weight_parse(rvalue, weight);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "IO weight '%s' invalid. Ignoring.", rvalue);
+ return 0;
+ }
+
+ return 0;
+}
+
+int config_parse_io_device_weight(
+ 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) {
+
+ _cleanup_free_ char *path = NULL;
+ CGroupIODeviceWeight *w;
+ CGroupContext *c = data;
+ const char *weight;
+ uint64_t u;
+ size_t n;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ if (isempty(rvalue)) {
+ while (c->io_device_weights)
+ cgroup_context_free_io_device_weight(c, c->io_device_weights);
+
+ return 0;
+ }
+
+ n = strcspn(rvalue, WHITESPACE);
+ weight = rvalue + n;
+ weight += strspn(weight, WHITESPACE);
+
+ if (isempty(weight)) {
+ log_syntax(unit, LOG_ERR, filename, line, 0, "Expected block device and device weight. Ignoring.");
+ return 0;
+ }
+
+ path = strndup(rvalue, n);
+ if (!path)
+ return log_oom();
+
+ if (!path_startswith(path, "/dev")) {
+ log_syntax(unit, LOG_ERR, filename, line, 0, "Invalid device node path '%s'. Ignoring.", path);
+ return 0;
+ }
+
+ r = cg_weight_parse(weight, &u);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "IO weight '%s' invalid. Ignoring.", weight);
+ return 0;
+ }
+
+ assert(u != CGROUP_WEIGHT_INVALID);
+
+ w = new0(CGroupIODeviceWeight, 1);
+ if (!w)
+ return log_oom();
+
+ w->path = path;
+ path = NULL;
+
+ w->weight = u;
+
+ LIST_PREPEND(device_weights, c->io_device_weights, w);
+ return 0;
+}
+
+int config_parse_io_limit(
+ 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) {
+
+ _cleanup_free_ char *path = NULL;
+ CGroupIODeviceLimit *l = NULL, *t;
+ CGroupContext *c = data;
+ CGroupIOLimitType type;
+ const char *limit;
+ uint64_t num;
+ size_t n;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+
+ type = cgroup_io_limit_type_from_string(lvalue);
+ assert(type >= 0);
+
+ if (isempty(rvalue)) {
+ LIST_FOREACH(device_limits, l, c->io_device_limits)
+ l->limits[type] = cgroup_io_limit_defaults[type];
+ return 0;
+ }
+
+ n = strcspn(rvalue, WHITESPACE);
+ limit = rvalue + n;
+ limit += strspn(limit, WHITESPACE);
+
+ if (!*limit) {
+ log_syntax(unit, LOG_ERR, filename, line, 0, "Expected space separated pair of device node and bandwidth. Ignoring.");
+ return 0;
+ }
+
+ path = strndup(rvalue, n);
+ if (!path)
+ return log_oom();
+
+ if (!path_startswith(path, "/dev")) {
+ log_syntax(unit, LOG_ERR, filename, line, 0, "Invalid device node path '%s'. Ignoring.", path);
+ return 0;
+ }
+
+ if (streq("max", limit)) {
+ num = CGROUP_LIMIT_MAX;
+ } else {
+ r = parse_size(limit, 1000, &num);
+ if (r < 0 || num <= 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r, "IO Limit '%s' invalid. Ignoring.", rvalue);
+ return 0;
+ }
+ }
+
+ LIST_FOREACH(device_limits, t, c->io_device_limits) {
+ if (path_equal(path, t->path)) {
+ l = t;
+ break;
+ }
+ }
+
+ if (!l) {
+ CGroupIOLimitType ttype;
+
+ l = new0(CGroupIODeviceLimit, 1);
+ if (!l)
+ return log_oom();
+
+ l->path = path;
+ path = NULL;
+ for (ttype = 0; ttype < _CGROUP_IO_LIMIT_TYPE_MAX; ttype++)
+ l->limits[ttype] = cgroup_io_limit_defaults[ttype];
+
+ LIST_PREPEND(device_limits, c->io_device_limits, l);
+ }
+
+ l->limits[type] = num;
+
+ return 0;
+}
+
int config_parse_blockio_weight(
const char *unit,
const char *filename,
@@ -3021,7 +3208,7 @@ int config_parse_blockio_bandwidth(
void *userdata) {
_cleanup_free_ char *path = NULL;
- CGroupBlockIODeviceBandwidth *b;
+ CGroupBlockIODeviceBandwidth *b = NULL, *t;
CGroupContext *c = data;
const char *bandwidth;
uint64_t bytes;
@@ -3036,12 +3223,10 @@ int config_parse_blockio_bandwidth(
read = streq("BlockIOReadBandwidth", lvalue);
if (isempty(rvalue)) {
- CGroupBlockIODeviceBandwidth *next;
-
- LIST_FOREACH_SAFE (device_bandwidths, b, next, c->blockio_device_bandwidths)
- if (b->read == read)
- cgroup_context_free_blockio_device_bandwidth(c, b);
-
+ LIST_FOREACH(device_bandwidths, b, c->blockio_device_bandwidths) {
+ b->rbps = CGROUP_LIMIT_MAX;
+ b->wbps = CGROUP_LIMIT_MAX;
+ }
return 0;
}
@@ -3069,16 +3254,30 @@ int config_parse_blockio_bandwidth(
return 0;
}
- b = new0(CGroupBlockIODeviceBandwidth, 1);
- if (!b)
- return log_oom();
+ LIST_FOREACH(device_bandwidths, t, c->blockio_device_bandwidths) {
+ if (path_equal(path, t->path)) {
+ b = t;
+ break;
+ }
+ }
- b->path = path;
- path = NULL;
- b->bandwidth = bytes;
- b->read = read;
+ if (!t) {
+ b = new0(CGroupBlockIODeviceBandwidth, 1);
+ if (!b)
+ return log_oom();
- LIST_PREPEND(device_bandwidths, c->blockio_device_bandwidths, b);
+ b->path = path;
+ path = NULL;
+ b->rbps = CGROUP_LIMIT_MAX;
+ b->wbps = CGROUP_LIMIT_MAX;
+
+ LIST_PREPEND(device_bandwidths, c->blockio_device_bandwidths, b);
+ }
+
+ if (read)
+ b->rbps = bytes;
+ else
+ b->wbps = bytes;
return 0;
}
@@ -3830,6 +4029,9 @@ void unit_dump_config_items(FILE *f) {
{ config_parse_memory_limit, "LIMIT" },
{ config_parse_device_allow, "DEVICE" },
{ config_parse_device_policy, "POLICY" },
+ { config_parse_io_limit, "LIMIT" },
+ { config_parse_io_weight, "WEIGHT" },
+ { config_parse_io_device_weight, "DEVICEWEIGHT" },
{ config_parse_blockio_bandwidth, "BANDWIDTH" },
{ config_parse_blockio_weight, "WEIGHT" },
{ config_parse_blockio_device_weight, "DEVICEWEIGHT" },
diff --git a/src/core/load-fragment.h b/src/core/load-fragment.h
index 34f15afa62..b36a2e3a02 100644
--- a/src/core/load-fragment.h
+++ b/src/core/load-fragment.h
@@ -86,6 +86,9 @@ int config_parse_memory_limit(const char *unit, const char *filename, unsigned l
int config_parse_tasks_max(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
int config_parse_device_policy(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
int config_parse_device_allow(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_io_weight(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_io_device_weight(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
+int config_parse_io_limit(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
int config_parse_blockio_weight(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
int config_parse_blockio_device_weight(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
int config_parse_blockio_bandwidth(const char *unit, const char *filename, unsigned line, const char *section, unsigned section_line, const char *lvalue, int ltype, const char *rvalue, void *data, void *userdata);
diff --git a/src/core/main.c b/src/core/main.c
index ed4d42c8cc..6397aadc73 100644
--- a/src/core/main.c
+++ b/src/core/main.c
@@ -122,6 +122,7 @@ static usec_t arg_default_timer_accuracy_usec = 1 * USEC_PER_MINUTE;
static Set* arg_syscall_archs = NULL;
static FILE* arg_serialization = NULL;
static bool arg_default_cpu_accounting = false;
+static bool arg_default_io_accounting = false;
static bool arg_default_blockio_accounting = false;
static bool arg_default_memory_accounting = false;
static bool arg_default_tasks_accounting = true;
@@ -691,6 +692,7 @@ static int parse_config_file(void) {
{ "Manager", "DefaultLimitRTPRIO", config_parse_limit, RLIMIT_RTPRIO, arg_default_rlimit },
{ "Manager", "DefaultLimitRTTIME", config_parse_limit, RLIMIT_RTTIME, arg_default_rlimit },
{ "Manager", "DefaultCPUAccounting", config_parse_bool, 0, &arg_default_cpu_accounting },
+ { "Manager", "DefaultIOAccounting", config_parse_bool, 0, &arg_default_io_accounting },
{ "Manager", "DefaultBlockIOAccounting", config_parse_bool, 0, &arg_default_blockio_accounting },
{ "Manager", "DefaultMemoryAccounting", config_parse_bool, 0, &arg_default_memory_accounting },
{ "Manager", "DefaultTasksAccounting", config_parse_bool, 0, &arg_default_tasks_accounting },
@@ -733,6 +735,7 @@ static void manager_set_defaults(Manager *m) {
m->default_start_limit_interval = arg_default_start_limit_interval;
m->default_start_limit_burst = arg_default_start_limit_burst;
m->default_cpu_accounting = arg_default_cpu_accounting;
+ m->default_io_accounting = arg_default_io_accounting;
m->default_blockio_accounting = arg_default_blockio_accounting;
m->default_memory_accounting = arg_default_memory_accounting;
m->default_tasks_accounting = arg_default_tasks_accounting;
diff --git a/src/core/manager.c b/src/core/manager.c
index e192cd475d..7838f56fd2 100644
--- a/src/core/manager.c
+++ b/src/core/manager.c
@@ -1494,7 +1494,7 @@ void manager_clear_jobs(Manager *m) {
while ((j = hashmap_first(m->jobs)))
/* No need to recurse. We're cancelling all jobs. */
- job_finish_and_invalidate(j, JOB_CANCELED, false);
+ job_finish_and_invalidate(j, JOB_CANCELED, false, false);
}
static int manager_dispatch_run_queue(sd_event_source *source, void *userdata) {
diff --git a/src/core/manager.h b/src/core/manager.h
index 4bccca75cb..6ed15c1a41 100644
--- a/src/core/manager.h
+++ b/src/core/manager.h
@@ -255,6 +255,7 @@ struct Manager {
bool default_cpu_accounting;
bool default_memory_accounting;
+ bool default_io_accounting;
bool default_blockio_accounting;
bool default_tasks_accounting;
diff --git a/src/core/namespace.c b/src/core/namespace.c
index b573f008b9..203d122810 100644
--- a/src/core/namespace.c
+++ b/src/core/namespace.c
@@ -44,6 +44,8 @@
#include "user-util.h"
#include "util.h"
+#define DEV_MOUNT_OPTIONS (MS_NOSUID|MS_STRICTATIME|MS_NOEXEC)
+
typedef enum MountMode {
/* This is ordered by priority! */
INACCESSIBLE,
@@ -153,7 +155,7 @@ static int mount_dev(BindMount *m) {
dev = strjoina(temporary_mount, "/dev");
(void) mkdir(dev, 0755);
- if (mount("tmpfs", dev, "tmpfs", MS_NOSUID|MS_STRICTATIME, "mode=755") < 0) {
+ if (mount("tmpfs", dev, "tmpfs", DEV_MOUNT_OPTIONS, "mode=755") < 0) {
r = -errno;
goto fail;
}
@@ -237,6 +239,8 @@ static int mount_dev(BindMount *m) {
*/
(void) mkdir_p_label(m->path, 0755);
+ /* Unmount everything in old /dev */
+ umount_recursive(m->path, 0);
if (mount(dev, m->path, NULL, MS_MOVE, NULL) < 0) {
r = -errno;
goto fail;
@@ -328,9 +332,11 @@ static int make_read_only(BindMount *m) {
if (IN_SET(m->mode, INACCESSIBLE, READONLY))
r = bind_remount_recursive(m->path, true);
- else if (IN_SET(m->mode, READWRITE, PRIVATE_TMP, PRIVATE_VAR_TMP, PRIVATE_DEV))
+ else if (IN_SET(m->mode, READWRITE, PRIVATE_TMP, PRIVATE_VAR_TMP, PRIVATE_DEV)) {
r = bind_remount_recursive(m->path, false);
- else
+ if (r == 0 && m->mode == PRIVATE_DEV) /* can be readonly but the submounts can't*/
+ r = mount(NULL, m->path, NULL, MS_REMOUNT|DEV_MOUNT_OPTIONS|MS_RDONLY, NULL);
+ } else
r = 0;
if (m->ignore && r == -ENOENT)
diff --git a/src/core/system.conf b/src/core/system.conf
index eacd7ee282..db8b7acd78 100644
--- a/src/core/system.conf
+++ b/src/core/system.conf
@@ -38,6 +38,7 @@
#DefaultStartLimitBurst=5
#DefaultEnvironment=
#DefaultCPUAccounting=no
+#DefaultIOAccounting=no
#DefaultBlockIOAccounting=no
#DefaultMemoryAccounting=no
#DefaultTasksAccounting=yes
diff --git a/src/core/transaction.c b/src/core/transaction.c
index d5370b2a14..e06a48a2f1 100644
--- a/src/core/transaction.c
+++ b/src/core/transaction.c
@@ -597,7 +597,7 @@ static int transaction_apply(Transaction *tr, Manager *m, JobMode mode) {
/* Not invalidating recursively. Avoids triggering
* OnFailure= actions of dependent jobs. Also avoids
* invalidating our iterator. */
- job_finish_and_invalidate(j, JOB_CANCELED, false);
+ job_finish_and_invalidate(j, JOB_CANCELED, false, false);
}
}
diff --git a/src/core/unit.c b/src/core/unit.c
index d8ab5781b0..2fff3f2d8b 100644
--- a/src/core/unit.c
+++ b/src/core/unit.c
@@ -132,6 +132,7 @@ static void unit_init(Unit *u) {
* been initialized */
cc->cpu_accounting = u->manager->default_cpu_accounting;
+ cc->io_accounting = u->manager->default_io_accounting;
cc->blockio_accounting = u->manager->default_blockio_accounting;
cc->memory_accounting = u->manager->default_memory_accounting;
cc->tasks_accounting = u->manager->default_tasks_accounting;
@@ -1213,6 +1214,7 @@ static int unit_add_startup_units(Unit *u) {
return 0;
if (c->startup_cpu_shares == CGROUP_CPU_SHARES_INVALID &&
+ c->startup_io_weight == CGROUP_WEIGHT_INVALID &&
c->startup_blockio_weight == CGROUP_BLKIO_WEIGHT_INVALID)
return 0;
@@ -1916,12 +1918,12 @@ void unit_notify(Unit *u, UnitActiveState os, UnitActiveState ns, bool reload_su
case JOB_VERIFY_ACTIVE:
if (UNIT_IS_ACTIVE_OR_RELOADING(ns))
- job_finish_and_invalidate(u->job, JOB_DONE, true);
+ job_finish_and_invalidate(u->job, JOB_DONE, true, false);
else if (u->job->state == JOB_RUNNING && ns != UNIT_ACTIVATING) {
unexpected = true;
if (UNIT_IS_INACTIVE_OR_FAILED(ns))
- job_finish_and_invalidate(u->job, ns == UNIT_FAILED ? JOB_FAILED : JOB_DONE, true);
+ job_finish_and_invalidate(u->job, ns == UNIT_FAILED ? JOB_FAILED : JOB_DONE, true, false);
}
break;
@@ -1932,12 +1934,12 @@ void unit_notify(Unit *u, UnitActiveState os, UnitActiveState ns, bool reload_su
if (u->job->state == JOB_RUNNING) {
if (ns == UNIT_ACTIVE)
- job_finish_and_invalidate(u->job, reload_success ? JOB_DONE : JOB_FAILED, true);
+ job_finish_and_invalidate(u->job, reload_success ? JOB_DONE : JOB_FAILED, true, false);
else if (ns != UNIT_ACTIVATING && ns != UNIT_RELOADING) {
unexpected = true;
if (UNIT_IS_INACTIVE_OR_FAILED(ns))
- job_finish_and_invalidate(u->job, ns == UNIT_FAILED ? JOB_FAILED : JOB_DONE, true);
+ job_finish_and_invalidate(u->job, ns == UNIT_FAILED ? JOB_FAILED : JOB_DONE, true, false);
}
}
@@ -1948,10 +1950,10 @@ void unit_notify(Unit *u, UnitActiveState os, UnitActiveState ns, bool reload_su
case JOB_TRY_RESTART:
if (UNIT_IS_INACTIVE_OR_FAILED(ns))
- job_finish_and_invalidate(u->job, JOB_DONE, true);
+ job_finish_and_invalidate(u->job, JOB_DONE, true, false);
else if (u->job->state == JOB_RUNNING && ns != UNIT_DEACTIVATING) {
unexpected = true;
- job_finish_and_invalidate(u->job, JOB_FAILED, true);
+ job_finish_and_invalidate(u->job, JOB_FAILED, true, false);
}
break;
@@ -2400,9 +2402,11 @@ int unit_set_slice(Unit *u, Unit *slice) {
if (UNIT_DEREF(u->slice) == slice)
return 0;
- if (UNIT_ISSET(u->slice))
+ /* Disallow slice changes if @u is already bound to cgroups */
+ if (UNIT_ISSET(u->slice) && u->cgroup_realized)
return -EBUSY;
+ unit_ref_unset(&u->slice);
unit_ref_set(&u->slice, slice);
return 1;
}
generated by cgit v1.2.3-54-g00ecf (git 2.45.2) at 2025-10-13 05:18:37 +0000