Re: [PATCH 1/1] netstress: Restore runtime to 5m

[Petr Vorel <pvorel@suse.cz>]

Hi,

> netstress requires the previous default timeout 5m due longer timeout
> for higher message sizes (e.g. 65535):

> ./sctp_ipsec.sh -6 -p comp -m transport -s 100:1000:65535:R65535
...
> sctp_ipsec 3 TINFO: run client 'netstress -l -T sctp -H fd00:1:1:1::1 -n 65535 -N 65535 -S fd00:1:1:1::2 -D ltp_ns_veth2 -a 2 -r 100 -d /tmp/LTP_sctp_ipsec.ARZbGkvjPa/tst_netload.res' 5 times
> sctp_ipsec 3 TWARN: netstress failed, ret: 2
> tst_test.c:1526: TINFO: Timeout per run is 0h 00m 30s
> netstress.c:896: TINFO: IP_BIND_ADDRESS_NO_PORT is used
> netstress.c:898: TINFO: connection: addr 'fd00:1:1:1::1', port '55097'
> netstress.c:900: TINFO: client max req: 100
> netstress.c:901: TINFO: clients num: 2
> netstress.c:906: TINFO: client msg size: 65535
> netstress.c:907: TINFO: server msg size: 65535
> netstress.c:979: TINFO: SCTP client
> netstress.c:475: TINFO: Running the test over IPv6
> Test timeouted, sending SIGKILL!
> tst_test.c:1577: TINFO: If you are running on slow machine, try exporting LTP_TIMEOUT_MUL > 1
> tst_test.c:1579: TBROK: Test killed! (timeout?)

> Converting netstress.c to use TST_NO_DEFAULT_MAIN (i.e. implementing main)
> would require more changes, because it uses .forks_child, .needs_checkpoints,
> cleanup function.
Merged to get at least netstress timeout fixed.

> NOTE: there is also needed to increase timeout for sctp_ipsec.sh.
> At least on my VM 10 min wasn't enough. Trying to measure reasonable
> time with TST_TIMEOUT=-1.
Continue debugging this.

Kind regards,
Petr