From: yufengmx <yufengx.mo@intel.com>
To: dts@dpdk.org, yinan.wang@intel.com
Cc: yufengmx <yufengx.mo@intel.com>
Subject: [dts] [PATCH V2 3/5] framework/pktgen_base: measure throughput supports multiple
Date: Fri, 22 Nov 2019 16:58:41 +0800 [thread overview]
Message-ID: <20191122085843.97671-4-yufengx.mo@intel.com> (raw)
In-Reply-To: <20191122085843.97671-1-yufengx.mo@intel.com>
return values
*. add test method(latency/loss/throughput) delay/duration options new usage
definition and relevant process source code.
*. set delay option to be the warm up time after start transmission.
*. add __get_single_throughput_statistic/__get_multi_throughput_statistic methods to realize
measure_throughput support return several throughput numbers in a duration.
*. set duration option default value to 10 second.
*. add test method(latency/loss/throughput/rfc2544) options parameter usage comment,
which is the same as doc pktgen_prog_guide.rst.
*. add a callback to deal with executing other tools query actions in throughput traffic lasting status.
Signed-off-by: yufengmx <yufengx.mo@intel.com>
---
framework/pktgen_base.py | 162 +++++++++++++++++++++++++++++++++++----
1 file changed, 148 insertions(+), 14 deletions(-)
diff --git a/framework/pktgen_base.py b/framework/pktgen_base.py
index 7855f6f..26f2f1a 100644
--- a/framework/pktgen_base.py
+++ b/framework/pktgen_base.py
@@ -157,39 +157,119 @@ class PacketGenerator(object):
def reset_streams(self):
self.__streams = []
- def measure_throughput(self, stream_ids=[], options={}):
- """
- Measure throughput on each tx ports
- """
- bps_rx = []
- pps_rx = []
- self._prepare_transmission(stream_ids=stream_ids)
+ def __warm_up_pktgen(self, stream_ids, options, delay):
+ ''' run warm up traffic before start main traffic '''
+ if not delay:
+ return
+ msg = '{1} packet generator: run traffic {0}s to warm up ... '.format(
+ delay, self.pktgen_type)
+ self.logger.info(msg)
self._start_transmission(stream_ids, options)
-
- delay = options.get('delay') or 5
time.sleep(delay)
+ self._stop_transmission(stream_ids)
+
+ def __get_single_throughput_statistic(self, stream_ids):
+ bps_rx = []
+ pps_rx = []
used_rx_port = []
+ msg = 'begin get port statistic ...'
+ self.logger.info(msg)
for stream_id in stream_ids:
if self.__streams[stream_id]['rx_port'] not in used_rx_port:
rxbps_rates, rxpps_rates = self._retrieve_port_statistic(
- stream_id, 'throughput')
+ stream_id, 'throughput')
used_rx_port.append(self.__streams[stream_id]['rx_port'])
bps_rx.append(rxbps_rates)
pps_rx.append(rxpps_rates)
- self._stop_transmission(stream_id)
-
bps_rx_total = self._summary_statistic(bps_rx)
pps_rx_total = self._summary_statistic(pps_rx)
- self.logger.info("throughput: pps_rx %f, bps_rx %f" % (pps_rx_total, bps_rx_total))
+ self.logger.info(
+ "throughput: pps_rx %f, bps_rx %f" % (pps_rx_total, bps_rx_total))
return bps_rx_total, pps_rx_total
+ def __get_multi_throughput_statistic(
+ self, stream_ids, duration, interval, callback=None):
+ """
+ duration: traffic duration (second)
+ interval: interval of get throughput statistics (second)
+ callback: a callback method of suite, which is used to do some actions
+ during traffic lasting.
+
+ Return: a list of throughput instead of a single tuple of pps/bps rate
+ """
+ time_elapsed = 0
+ stats = []
+ while time_elapsed < duration:
+ time.sleep(interval)
+ stats.append(self.__get_single_throughput_statistic(stream_ids))
+ if callback and callable(callback):
+ callback()
+ time_elapsed += interval
+ return stats
+
+ def measure_throughput(self, stream_ids=[], options={}):
+ """
+ Measure throughput on each tx ports
+
+ options usage:
+ rate:
+ port rate percent, float(0--100). Default value is 100.
+
+ delay:
+ warm up time before start main traffic. If it is set, it will start
+ a delay time traffic to make sure packet generator under good status.
+ Warm up flow is ignored by default.
+
+ interval:
+ a interval time of get throughput statistic (second)
+ If set this key value, pktgen will return several throughput statistic
+ data within a duration traffic. If not set this key value, only
+ return one statistic data. It is ignored by default.
+
+ callback:
+ this key works with ``interval`` key. If it is set, the callback
+ of suite level will be executed after getting throughput statistic.
+ callback method should define as below, don't add sleep in this method.
+
+ def callback(self):
+ xxxx()
+
+ duration:
+ traffic lasting time(second). Default value is 10 second.
+ """
+ interval = options.get('interval')
+ callback = options.get('callback')
+ duration = options.get('duration') or 10
+ delay = options.get('delay')
+ self._prepare_transmission(stream_ids=stream_ids)
+ # start warm up traffic
+ self.__warm_up_pktgen(stream_ids, options, delay)
+ # main traffic
+ self._start_transmission(stream_ids)
+ # keep traffic within a duration time and get throughput statistic
+ if interval and duration:
+ stats = self.__get_multi_throughput_statistic(
+ stream_ids, duration, interval, callback)
+ else:
+ time.sleep(duration)
+ stats = self.__get_single_throughput_statistic(stream_ids)
+ self._stop_transmission(stream_ids)
+ return stats
+
def _measure_loss(self, stream_ids=[], options={}):
"""
Measure lost rate on each tx/rx ports
"""
+ delay = options.get('delay')
+ duration = options.get('duration') or 10
self._prepare_transmission(stream_ids=stream_ids)
+ # start warm up traffic
+ self.__warm_up_pktgen(stream_ids, options, delay)
+ # main traffic
self._start_transmission(stream_ids, options)
+ # keep traffic within a duration time
+ time.sleep(duration)
self._stop_transmission(None)
result = {}
used_rx_port = []
@@ -210,6 +290,19 @@ class PacketGenerator(object):
return result
def measure_loss(self, stream_ids=[], options={}):
+ '''
+ options usage:
+ rate:
+ port rate percent, float(0--100). Default value is 100.
+
+ delay:
+ warm up time before start main traffic. If it is set, it will
+ start a delay time traffic to make sure packet generator
+ under good status. Warm up flow is ignored by default.
+
+ duration:
+ traffic lasting time(second). Default value is 10 second.
+ '''
result = self._measure_loss(stream_ids, options)
# here only to make sure that return value is the same as dts/etgen format
# In real testing scenario, this method can offer more data than it
@@ -218,9 +311,28 @@ class PacketGenerator(object):
def measure_latency(self, stream_ids=[], options={}):
"""
Measure latency on each tx/rx ports
+
+ options usage:
+ rate:
+ port rate percent, float(0--100). Default value is 100.
+
+ delay:
+ warm up time before start main traffic. If it is set, it will
+ start a delay time transmission to make sure packet generator
+ under correct status. Warm up flow is ignored by default.
+
+ duration:
+ traffic lasting time(second). Default value is 10 second.
"""
+ delay = options.get('delay')
+ duration = options.get('duration') or 10
self._prepare_transmission(stream_ids=stream_ids, latency=True)
+ # start warm up traffic
+ self.__warm_up_pktgen(stream_ids, options, delay)
+ # main traffic
self._start_transmission(stream_ids, options)
+ # keep traffic within a duration time
+ time.sleep(duration)
self._stop_transmission(None)
result = {}
@@ -248,7 +360,26 @@ class PacketGenerator(object):
return True
def measure_rfc2544(self, stream_ids=[], options={}):
- """ check loss rate with rate percent dropping """
+ """ check loss rate with rate percent dropping
+
+ options usage:
+ rate:
+ port rate percent at first round testing(0 ~ 100), default is 100.
+
+ pdr:
+ permit packet drop rate, , default is 0.
+
+ drop_step:
+ port rate percent drop step(0 ~ 100), default is 1.
+
+ delay:
+ warm up time before start main traffic. If it is set, it will
+ start a delay time traffic to make sure packet generator
+ under good status. Warm up flow is ignored by default.
+
+ duration:
+ traffic lasting time(second). Default value is 10 second.
+ """
loss_rate_table = []
rate_percent = options.get('rate') or float(100)
permit_loss_rate = options.get('pdr') or 0
@@ -265,6 +396,9 @@ class PacketGenerator(object):
tx_num, rx_num = result.values()[0][1:]
return rate_percent, tx_num, rx_num
_options = deepcopy(options)
+ # if warm up option 'delay' is set, ignore it in next work flow
+ if 'delay' in _options:
+ _options.pop('delay')
if 'rate' in _options:
_options.pop('rate')
while not status and rate_percent > 0:
--
2.21.0
next prev parent reply other threads:[~2019-11-22 8:57 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-22 8:58 [dts] [PATCH V2 0/5] dts/pktgen: measure throughput supports multiple return values and callback yufengmx
2019-11-22 8:58 ` [dts] [PATCH V2 1/5] conf/pktgen: remove un-used optons yufengmx
2019-11-22 8:58 ` [dts] [PATCH V2 2/5] doc/dts_gsg/pktgen_prog_guide: update description yufengmx
2019-11-22 8:58 ` yufengmx [this message]
2019-11-22 8:58 ` [dts] [PATCH V2 4/5] framework/pktgen_ixia: measure throughput supports multiple yufengmx
2019-11-22 8:58 ` [dts] [PATCH V2 5/5] framework/pktgen_trex: " yufengmx
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20191122085843.97671-4-yufengx.mo@intel.com \
--to=yufengx.mo@intel.com \
--cc=dts@dpdk.org \
--cc=yinan.wang@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).