diff options
Diffstat (limited to 'tools/testing/selftests/drivers/net/stats.py')
| -rwxr-xr-x | tools/testing/selftests/drivers/net/stats.py | 137 |
1 files changed, 120 insertions, 17 deletions
diff --git a/tools/testing/selftests/drivers/net/stats.py b/tools/testing/selftests/drivers/net/stats.py index 031ac9def6c0..c2bb5d3f1ca1 100755 --- a/tools/testing/selftests/drivers/net/stats.py +++ b/tools/testing/selftests/drivers/net/stats.py @@ -1,13 +1,20 @@ #!/usr/bin/env python3 # SPDX-License-Identifier: GPL-2.0 +""" +Tests related to standard netdevice statistics. +""" + import errno +import subprocess +import time from lib.py import ksft_run, ksft_exit, ksft_pr -from lib.py import ksft_ge, ksft_eq, ksft_in, ksft_true, ksft_raises, KsftSkipEx, KsftXfailEx +from lib.py import ksft_ge, ksft_eq, ksft_is, ksft_in, ksft_lt, ksft_true, ksft_raises +from lib.py import KsftSkipEx, KsftFailEx from lib.py import ksft_disruptive from lib.py import EthtoolFamily, NetdevFamily, RtnlFamily, NlError from lib.py import NetDrvEnv -from lib.py import ip, defer +from lib.py import cmd, ip, defer ethnl = EthtoolFamily() netfam = NetdevFamily() @@ -15,13 +22,16 @@ rtnl = RtnlFamily() def check_pause(cfg) -> None: - global ethnl + """ + Check that drivers which support Pause config also report standard + pause stats. + """ try: ethnl.pause_get({"header": {"dev-index": cfg.ifindex}}) except NlError as e: if e.error == errno.EOPNOTSUPP: - raise KsftXfailEx("pause not supported by the device") + raise KsftSkipEx("pause not supported by the device") from e raise data = ethnl.pause_get({"header": {"dev-index": cfg.ifindex, @@ -30,13 +40,16 @@ def check_pause(cfg) -> None: def check_fec(cfg) -> None: - global ethnl + """ + Check that drivers which support FEC config also report standard + FEC stats. + """ try: ethnl.fec_get({"header": {"dev-index": cfg.ifindex}}) except NlError as e: if e.error == errno.EOPNOTSUPP: - raise KsftXfailEx("FEC not supported by the device") + raise KsftSkipEx("FEC not supported by the device") from e raise data = ethnl.fec_get({"header": {"dev-index": cfg.ifindex, @@ -45,15 +58,17 @@ def check_fec(cfg) -> None: def pkt_byte_sum(cfg) -> None: - global netfam, rtnl + """ + Check that qstat and interface stats match in value. + """ def get_qstat(test): - global netfam stats = netfam.qstats_get({}, dump=True) if stats: for qs in stats: if qs["ifindex"]== test.ifindex: return qs + return None qstat = get_qstat(cfg) if qstat is None: @@ -74,15 +89,14 @@ def pkt_byte_sum(cfg) -> None: for _ in range(10): rtstat = rtnl.getlink({"ifi-index": cfg.ifindex})['stats64'] if stat_cmp(rtstat, qstat) < 0: - raise Exception("RTNL stats are lower, fetched later") + raise KsftFailEx("RTNL stats are lower, fetched later") qstat = get_qstat(cfg) if stat_cmp(rtstat, qstat) > 0: - raise Exception("Qstats are lower, fetched later") + raise KsftFailEx("Qstats are lower, fetched later") def qstat_by_ifindex(cfg) -> None: - global netfam - global rtnl + """ Qstats Netlink API tests - querying by ifindex. """ # Construct a map ifindex -> [dump, by-index, dump] ifindexes = {} @@ -90,7 +104,7 @@ def qstat_by_ifindex(cfg) -> None: for entry in stats: ifindexes[entry['ifindex']] = [entry, None, None] - for ifindex in ifindexes.keys(): + for ifindex in ifindexes: entry = netfam.qstats_get({"ifindex": ifindex}, dump=True) ksft_eq(len(entry), 1) ifindexes[entry[0]['ifindex']][1] = entry[0] @@ -142,7 +156,7 @@ def qstat_by_ifindex(cfg) -> None: # Try to get stats for lowest unused ifindex but not 0 devs = rtnl.getlink({}, dump=True) - all_ifindexes = set([dev["ifi-index"] for dev in devs]) + all_ifindexes = set(dev["ifi-index"] for dev in devs) lowest = 2 while lowest in all_ifindexes: lowest += 1 @@ -155,18 +169,20 @@ def qstat_by_ifindex(cfg) -> None: @ksft_disruptive def check_down(cfg) -> None: + """ Test statistics (interface and qstat) are not impacted by ifdown """ + try: qstat = netfam.qstats_get({"ifindex": cfg.ifindex}, dump=True)[0] except NlError as e: if e.error == errno.EOPNOTSUPP: - raise KsftSkipEx("qstats not supported by the device") + raise KsftSkipEx("qstats not supported by the device") from e raise ip(f"link set dev {cfg.dev['ifname']} down") defer(ip, f"link set dev {cfg.dev['ifname']} up") qstat2 = netfam.qstats_get({"ifindex": cfg.ifindex}, dump=True)[0] - for k, v in qstat.items(): + for k in qstat: ksft_ge(qstat2[k], qstat[k], comment=f"{k} went backwards on device down") # exercise per-queue API to make sure that "device down" state @@ -174,10 +190,97 @@ def check_down(cfg) -> None: netfam.qstats_get({"ifindex": cfg.ifindex, "scope": "queue"}, dump=True) +def __run_inf_loop(body): + body = body.strip() + if body[-1] != ';': + body += ';' + + return subprocess.Popen(f"while true; do {body} done", shell=True, + stdout=subprocess.PIPE, stderr=subprocess.PIPE) + + +def __stats_increase_sanely(old, new) -> None: + for k in old.keys(): + ksft_ge(new[k], old[k]) + ksft_lt(new[k] - old[k], 1 << 31, comment="likely wrapping error") + + +def procfs_hammer(cfg) -> None: + """ + Reading stats via procfs only holds the RCU lock, which is not an exclusive + lock, make sure drivers can handle parallel reads of stats. + """ + one = __run_inf_loop("cat /proc/net/dev") + defer(one.kill) + two = __run_inf_loop("cat /proc/net/dev") + defer(two.kill) + + time.sleep(1) + # Make sure the processes are running + ksft_is(one.poll(), None) + ksft_is(two.poll(), None) + + rtstat1 = rtnl.getlink({"ifi-index": cfg.ifindex})['stats64'] + time.sleep(2) + rtstat2 = rtnl.getlink({"ifi-index": cfg.ifindex})['stats64'] + __stats_increase_sanely(rtstat1, rtstat2) + # defers will kill the loops + + +@ksft_disruptive +def procfs_downup_hammer(cfg) -> None: + """ + Reading stats via procfs only holds the RCU lock, drivers often try + to sleep when reading the stats, or don't protect against races. + """ + # Max out the queues, we'll flip between max and 1 + channels = ethnl.channels_get({'header': {'dev-index': cfg.ifindex}}) + if channels['combined-count'] == 0: + rx_type = 'rx' + else: + rx_type = 'combined' + cur_queue_cnt = channels[f'{rx_type}-count'] + max_queue_cnt = channels[f'{rx_type}-max'] + + cmd(f"ethtool -L {cfg.ifname} {rx_type} {max_queue_cnt}") + defer(cmd, f"ethtool -L {cfg.ifname} {rx_type} {cur_queue_cnt}") + + # Real test stats + stats = __run_inf_loop("cat /proc/net/dev") + defer(stats.kill) + + ipset = f"ip link set dev {cfg.ifname}" + defer(ip, f"link set dev {cfg.ifname} up") + # The "echo -n 1" lets us count iterations below + updown = f"{ipset} down; sleep 0.05; {ipset} up; sleep 0.05; " + \ + f"ethtool -L {cfg.ifname} {rx_type} 1; " + \ + f"ethtool -L {cfg.ifname} {rx_type} {max_queue_cnt}; " + \ + "echo -n 1" + updown = __run_inf_loop(updown) + kill_updown = defer(updown.kill) + + time.sleep(1) + # Make sure the processes are running + ksft_is(stats.poll(), None) + ksft_is(updown.poll(), None) + + rtstat1 = rtnl.getlink({"ifi-index": cfg.ifindex})['stats64'] + # We're looking for crashes, give it extra time + time.sleep(9) + rtstat2 = rtnl.getlink({"ifi-index": cfg.ifindex})['stats64'] + __stats_increase_sanely(rtstat1, rtstat2) + + kill_updown.exec() + stdout, _ = updown.communicate(timeout=5) + ksft_pr("completed up/down cycles:", len(stdout.decode('utf-8'))) + + def main() -> None: + """ Ksft boiler plate main """ + with NetDrvEnv(__file__, queue_count=100) as cfg: ksft_run([check_pause, check_fec, pkt_byte_sum, qstat_by_ifindex, - check_down], + check_down, procfs_hammer, procfs_downup_hammer], args=(cfg, )) ksft_exit() |