</listitem>
</varlistentry>
+ <varlistentry>
+ <term><varname>FairQueueTrafficPolicingFlowLimit=</varname></term>
+ <listitem>
+ <para>Specifies the hard limit on the maximum number of packets queued per flow. Defaults to
+ unset and kernel's default is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>FairQueueTrafficPolicingQuantum=</varname></term>
+ <listitem>
+ <para>Specifies the credit per dequeue RR round, i.e. the amount of bytes a flow is allowed
+ to dequeue at once. When suffixed with K, M, or G, the specified size is parsed as Kilobytes,
+ Megabytes, or Gigabytes, respectively, to the base of 1024. Defaults to unset and kernel's
+ default is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>FairQueueTrafficPolicingInitialQuantum=</varname></term>
+ <listitem>
+ <para>Specifies the initial sending rate credit, i.e. the amount of bytes a new flow is
+ allowed to dequeue initially. When suffixed with K, M, or G, the specified size is parsed as
+ Kilobytes, Megabytes, or Gigabytes, respectively, to the base of 1024. Defaults to unset and
+ kernel's default is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>FairQueueTrafficPolicingMaximumRate=</varname></term>
+ <listitem>
+ <para>Specifies the maximum sending rate of a flow. When suffixed with K, M, or G, the
+ specified size is parsed as Kilobytes, Megabytes, or Gigabytes, respectively, to the base of
+ 1000. Defaults to unset and kernel's default is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>FairQueueTrafficPolicingBuckets=</varname></term>
+ <listitem>
+ <para>Specifies the size of the hash table used for flow lookups. Defaults to unset and
+ kernel's default is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>FairQueueTrafficPolicingOrphanMask=</varname></term>
+ <listitem>
+ <para>Takes an unsigned integer. For packets not owned by a socket, fq is able to mask a part
+ of hash and reduce number of buckets associated with the traffic. Defaults to unset and
+ kernel's default is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>FairQueueTrafficPolicingPacing=</varname></term>
+ <listitem>
+ <para>Takes a boolean, and enables or disables flow pacing. Defaults to unset and kernel's
+ default is used.</para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>FairQueueTrafficPolicingCEThresholdSec=</varname></term>
+ <listitem>
+ <para>Takes a timespan. This sets a threshold above which all packets are marked with ECN
+ Congestion Experienced (CE). Defaults to unset and kernel's default is used.</para>
+ </listitem>
+ </varlistentry>
+
</variablelist>
</refsect1>
TrafficControlQueueingDiscipline.FairQueuingControlledDelayIntervalSec, config_parse_tc_fair_queuing_controlled_delay_usec, 0, 0
TrafficControlQueueingDiscipline.FairQueuingControlledDelayCEThresholdSec, config_parse_tc_fair_queuing_controlled_delay_usec, 0, 0
TrafficControlQueueingDiscipline.FairQueuingControlledDelayECN, config_parse_tc_fair_queuing_controlled_delay_bool, 0, 0
-TrafficControlQueueingDiscipline.FairQueueTrafficPolicingPacketLimit, config_parse_tc_fair_queue_traffic_policing_packet_limit, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingPacketLimit, config_parse_tc_fair_queue_traffic_policing_u32, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingFlowLimit, config_parse_tc_fair_queue_traffic_policing_u32, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingQuantum, config_parse_tc_fair_queue_traffic_policing_size, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingInitialQuantum, config_parse_tc_fair_queue_traffic_policing_size, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingMaximumRate, config_parse_tc_fair_queue_traffic_policing_max_rate, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingBuckets, config_parse_tc_fair_queue_traffic_policing_u32, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingOrphanMask, config_parse_tc_fair_queue_traffic_policing_u32, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingPacing, config_parse_tc_fair_queue_traffic_policing_bool, 0, 0
+TrafficControlQueueingDiscipline.FairQueueTrafficPolicingCEThresholdSec, config_parse_tc_fair_queue_traffic_policing_usec, 0, 0
/* backwards compatibility: do not add new entries to this section */
Network.IPv4LL, config_parse_ipv4ll, 0, offsetof(Network, link_local)
DHCP.ClientIdentifier, config_parse_dhcp_client_identifier, 0, offsetof(Network, dhcp_client_identifier)
#include "netlink-util.h"
#include "parse-util.h"
#include "string-util.h"
+#include "util.h"
+
+static int fair_queue_traffic_policing_init(QDisc *qdisc) {
+ FairQueueTrafficPolicing *fq;
+
+ assert(qdisc);
+
+ fq = FQ(qdisc);
+
+ fq->pacing = -1;
+ fq->ce_threshold_usec = USEC_INFINITY;
+
+ return 0;
+}
static int fair_queue_traffic_policing_fill_message(Link *link, QDisc *qdisc, sd_netlink_message *req) {
FairQueueTrafficPolicing *fq;
if (r < 0)
return log_link_error_errno(link, r, "Could not open container TCA_OPTIONS: %m");
- r = sd_netlink_message_append_u32(req, TCA_FQ_PLIMIT, fq->limit);
- if (r < 0)
- return log_link_error_errno(link, r, "Could not append TCA_FQ_PLIMIT attribute: %m");
+ if (fq->packet_limit > 0) {
+ r = sd_netlink_message_append_u32(req, TCA_FQ_PLIMIT, fq->packet_limit);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_PLIMIT attribute: %m");
+ }
+
+ if (fq->flow_limit > 0) {
+ r = sd_netlink_message_append_u32(req, TCA_FQ_FLOW_PLIMIT, fq->flow_limit);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_FLOW_PLIMIT attribute: %m");
+ }
+
+ if (fq->quantum > 0) {
+ r = sd_netlink_message_append_u32(req, TCA_FQ_QUANTUM, fq->quantum);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_QUANTUM attribute: %m");
+ }
+
+ if (fq->initial_quantum > 0) {
+ r = sd_netlink_message_append_u32(req, TCA_FQ_INITIAL_QUANTUM, fq->initial_quantum);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_INITIAL_QUANTUM attribute: %m");
+ }
+
+ if (fq->pacing >= 0) {
+ r = sd_netlink_message_append_u32(req, TCA_FQ_RATE_ENABLE, fq->pacing);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_RATE_ENABLE attribute: %m");
+ }
+
+ if (fq->max_rate > 0) {
+ r = sd_netlink_message_append_u32(req, TCA_FQ_FLOW_MAX_RATE, fq->max_rate);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_FLOW_MAX_RATE attribute: %m");
+ }
+
+ if (fq->buckets > 0) {
+ uint32_t l;
+
+ l = log2u(fq->buckets);
+ r = sd_netlink_message_append_u32(req, TCA_FQ_BUCKETS_LOG, l);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_BUCKETS_LOG attribute: %m");
+ }
+
+ if (fq->orphan_mask > 0) {
+ r = sd_netlink_message_append_u32(req, TCA_FQ_ORPHAN_MASK, fq->orphan_mask);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_ORPHAN_MASK attribute: %m");
+ }
+
+ if (fq->ce_threshold_usec != USEC_INFINITY) {
+ r = sd_netlink_message_append_u32(req, TCA_FQ_CE_THRESHOLD, fq->ce_threshold_usec);
+ if (r < 0)
+ return log_link_error_errno(link, r, "Could not append TCA_FQ_CE_THRESHOLD attribute: %m");
+ }
r = sd_netlink_message_close_container(req);
if (r < 0)
return 0;
}
-int config_parse_tc_fair_queue_traffic_policing_packet_limit(
+int config_parse_tc_fair_queue_traffic_policing_u32(
const char *unit,
const char *filename,
unsigned line,
_cleanup_(qdisc_free_or_set_invalidp) QDisc *qdisc = NULL;
FairQueueTrafficPolicing *fq;
Network *network = data;
+ uint32_t *p;
int r;
assert(filename);
fq = FQ(qdisc);
+ if (streq(lvalue, "FairQueueTrafficPolicingPacketLimit"))
+ p = &fq->packet_limit;
+ else if (streq(lvalue, "FairQueueTrafficPolicingFlowLimit"))
+ p = &fq->flow_limit;
+ else if (streq(lvalue, "FairQueueTrafficPolicingBuckets"))
+ p = &fq->buckets;
+ else if (streq(lvalue, "FairQueueTrafficPolicingOrphanMask"))
+ p = &fq->orphan_mask;
+ else
+ assert_not_reached("Invalid lvalue");
+
if (isempty(rvalue)) {
- fq->limit = 0;
+ *p = 0;
qdisc = NULL;
return 0;
}
- r = safe_atou32(rvalue, &fq->limit);
+ r = safe_atou32(rvalue, p);
if (r < 0) {
log_syntax(unit, LOG_ERR, filename, line, r,
"Failed to parse '%s=', ignoring assignment: %s",
return 0;
}
+int config_parse_tc_fair_queue_traffic_policing_size(
+ 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_(qdisc_free_or_set_invalidp) QDisc *qdisc = NULL;
+ FairQueueTrafficPolicing *fq;
+ Network *network = data;
+ uint64_t sz;
+ uint32_t *p;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+ assert(data);
+
+ r = qdisc_new_static(QDISC_KIND_FQ, network, filename, section_line, &qdisc);
+ if (r == -ENOMEM)
+ return log_oom();
+ if (r < 0)
+ return log_syntax(unit, LOG_ERR, filename, line, r,
+ "More than one kind of queueing discipline, ignoring assignment: %m");
+
+ fq = FQ(qdisc);
+
+ if (streq(lvalue, "FairQueueTrafficPolicingQuantum"))
+ p = &fq->quantum;
+ else if (streq(lvalue, "FairQueueTrafficPolicingInitialQuantum"))
+ p = &fq->initial_quantum;
+ else
+ assert_not_reached("Invalid lvalue");
+
+ if (isempty(rvalue)) {
+ *p = 0;
+
+ qdisc = NULL;
+ return 0;
+ }
+
+ r = parse_size(rvalue, 1024, &sz);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r,
+ "Failed to parse '%s=', ignoring assignment: %s",
+ lvalue, rvalue);
+ return 0;
+ }
+ if (sz > UINT32_MAX) {
+ log_syntax(unit, LOG_ERR, filename, line, r,
+ "Specified '%s=' is too large, ignoring assignment: %s",
+ lvalue, rvalue);
+ return 0;
+ }
+
+ *p = sz;
+ qdisc = NULL;
+
+ return 0;
+}
+
+int config_parse_tc_fair_queue_traffic_policing_bool(
+ 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_(qdisc_free_or_set_invalidp) QDisc *qdisc = NULL;
+ FairQueueTrafficPolicing *fq;
+ Network *network = data;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+ assert(data);
+
+ r = qdisc_new_static(QDISC_KIND_FQ, network, filename, section_line, &qdisc);
+ if (r == -ENOMEM)
+ return log_oom();
+ if (r < 0)
+ return log_syntax(unit, LOG_ERR, filename, line, r,
+ "More than one kind of queueing discipline, ignoring assignment: %m");
+
+ fq = FQ(qdisc);
+
+ if (isempty(rvalue)) {
+ fq->pacing = -1;
+
+ qdisc = NULL;
+ return 0;
+ }
+
+ r = parse_boolean(rvalue);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r,
+ "Failed to parse '%s=', ignoring assignment: %s",
+ lvalue, rvalue);
+ return 0;
+ }
+
+ fq->pacing = r;
+ qdisc = NULL;
+
+ return 0;
+}
+
+int config_parse_tc_fair_queue_traffic_policing_usec(
+ 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_(qdisc_free_or_set_invalidp) QDisc *qdisc = NULL;
+ FairQueueTrafficPolicing *fq;
+ Network *network = data;
+ usec_t sec;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+ assert(data);
+
+ r = qdisc_new_static(QDISC_KIND_FQ, network, filename, section_line, &qdisc);
+ if (r == -ENOMEM)
+ return log_oom();
+ if (r < 0)
+ return log_syntax(unit, LOG_ERR, filename, line, r,
+ "More than one kind of queueing discipline, ignoring assignment: %m");
+
+ fq = FQ(qdisc);
+
+ if (isempty(rvalue)) {
+ fq->ce_threshold_usec = USEC_INFINITY;
+
+ qdisc = NULL;
+ return 0;
+ }
+
+ r = parse_sec(rvalue, &sec);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r,
+ "Failed to parse '%s=', ignoring assignment: %s",
+ lvalue, rvalue);
+ return 0;
+ }
+ if (sec > UINT32_MAX) {
+ log_syntax(unit, LOG_ERR, filename, line, 0,
+ "Specified '%s=' is too large, ignoring assignment: %s",
+ lvalue, rvalue);
+ return 0;
+ }
+
+ fq->ce_threshold_usec = sec;
+ qdisc = NULL;
+
+ return 0;
+}
+
+int config_parse_tc_fair_queue_traffic_policing_max_rate(
+ 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_(qdisc_free_or_set_invalidp) QDisc *qdisc = NULL;
+ FairQueueTrafficPolicing *fq;
+ Network *network = data;
+ uint64_t sz;
+ int r;
+
+ assert(filename);
+ assert(lvalue);
+ assert(rvalue);
+ assert(data);
+
+ r = qdisc_new_static(QDISC_KIND_FQ, network, filename, section_line, &qdisc);
+ if (r == -ENOMEM)
+ return log_oom();
+ if (r < 0)
+ return log_syntax(unit, LOG_ERR, filename, line, r,
+ "More than one kind of queueing discipline, ignoring assignment: %m");
+
+ fq = FQ(qdisc);
+
+ if (isempty(rvalue)) {
+ fq->max_rate = 0;
+
+ qdisc = NULL;
+ return 0;
+ }
+
+ r = parse_size(rvalue, 1000, &sz);
+ if (r < 0) {
+ log_syntax(unit, LOG_ERR, filename, line, r,
+ "Failed to parse '%s=', ignoring assignment: %s",
+ lvalue, rvalue);
+ return 0;
+ }
+ if (sz / 8 > UINT32_MAX) {
+ log_syntax(unit, LOG_ERR, filename, line, 0,
+ "Specified '%s=' is too large, ignoring assignment: %s",
+ lvalue, rvalue);
+ return 0;
+ }
+
+ fq->max_rate = sz / 8;
+ qdisc = NULL;
+
+ return 0;
+}
+
const QDiscVTable fq_vtable = {
+ .init = fair_queue_traffic_policing_init,
.object_size = sizeof(FairQueueTrafficPolicing),
.tca_kind = "fq",
.fill_message = fair_queue_traffic_policing_fill_message,
projects / thirdparty / systemd.git / commitdiff
