Modem sleep does not work

[readmodifywrite]

NONOS SDK 3.0.3

Calling wifi_set_sleep_type( MODEM_SLEEP_T ) appears to have no effect on the sleep state of the chip. My AP has a DTIM of 1, so I should be seeing ping times vary randomly to over 100 ms. However, the ping times I get are this:

64 bytes from 10.0.0.65: icmp_seq=1 ttl=128 time=6.28 ms
64 bytes from 10.0.0.65: icmp_seq=2 ttl=128 time=1.44 ms
64 bytes from 10.0.0.65: icmp_seq=3 ttl=128 time=1.98 ms
64 bytes from 10.0.0.65: icmp_seq=4 ttl=128 time=1.86 ms
64 bytes from 10.0.0.65: icmp_seq=5 ttl=128 time=2.71 ms
64 bytes from 10.0.0.65: icmp_seq=6 ttl=128 time=2.14 ms
64 bytes from 10.0.0.65: icmp_seq=7 ttl=128 time=1.99 ms
64 bytes from 10.0.0.65: icmp_seq=8 ttl=128 time=1.27 ms
64 bytes from 10.0.0.65: icmp_seq=9 ttl=128 time=1.52 ms
64 bytes from 10.0.0.65: icmp_seq=10 ttl=128 time=2.29 ms
64 bytes from 10.0.0.65: icmp_seq=11 ttl=128 time=3.78 ms
64 bytes from 10.0.0.65: icmp_seq=12 ttl=128 time=4.40 ms
64 bytes from 10.0.0.65: icmp_seq=13 ttl=128 time=1.44 ms
64 bytes from 10.0.0.65: icmp_seq=14 ttl=128 time=3.01 ms
64 bytes from 10.0.0.65: icmp_seq=15 ttl=128 time=1.82 ms
64 bytes from 10.0.0.65: icmp_seq=16 ttl=128 time=13.8 ms
64 bytes from 10.0.0.65: icmp_seq=17 ttl=128 time=2.48 ms
64 bytes from 10.0.0.65: icmp_seq=18 ttl=128 time=4.74 ms
64 bytes from 10.0.0.65: icmp_seq=19 ttl=128 time=2.20 ms
64 bytes from 10.0.0.65: icmp_seq=20 ttl=128 time=13.8 ms
64 bytes from 10.0.0.65: icmp_seq=21 ttl=128 time=2.87 ms
64 bytes from 10.0.0.65: icmp_seq=22 ttl=128 time=2.51 ms
64 bytes from 10.0.0.65: icmp_seq=23 ttl=128 time=1.91 ms
64 bytes from 10.0.0.65: icmp_seq=24 ttl=128 time=1.37 ms
64 bytes from 10.0.0.65: icmp_seq=25 ttl=128 time=1.62 ms
64 bytes from 10.0.0.65: icmp_seq=26 ttl=128 time=1.67 ms
64 bytes from 10.0.0.65: icmp_seq=27 ttl=128 time=2.48 ms
64 bytes from 10.0.0.65: icmp_seq=28 ttl=128 time=2.30 ms
64 bytes from 10.0.0.65: icmp_seq=29 ttl=128 time=2.61 ms
64 bytes from 10.0.0.65: icmp_seq=30 ttl=128 time=7.82 ms
64 bytes from 10.0.0.65: icmp_seq=31 ttl=128 time=2.12 ms
64 bytes from 10.0.0.65: icmp_seq=32 ttl=128 time=1.65 ms
64 bytes from 10.0.0.65: icmp_seq=33 ttl=128 time=2.83 ms
64 bytes from 10.0.0.65: icmp_seq=34 ttl=128 time=1.86 ms
64 bytes from 10.0.0.65: icmp_seq=35 ttl=128 time=2.11 ms
64 bytes from 10.0.0.65: icmp_seq=36 ttl=128 time=28.5 ms
64 bytes from 10.0.0.65: icmp_seq=37 ttl=128 time=11.8 ms
64 bytes from 10.0.0.65: icmp_seq=38 ttl=128 time=4.46 ms
64 bytes from 10.0.0.65: icmp_seq=39 ttl=128 time=2.71 ms
64 bytes from 10.0.0.65: icmp_seq=40 ttl=128 time=2.50 ms
64 bytes from 10.0.0.65: icmp_seq=41 ttl=128 time=2.33 ms
64 bytes from 10.0.0.65: icmp_seq=42 ttl=128 time=1.78 ms
64 bytes from 10.0.0.65: icmp_seq=43 ttl=128 time=6.18 ms
64 bytes from 10.0.0.65: icmp_seq=44 ttl=128 time=2.77 ms
64 bytes from 10.0.0.65: icmp_seq=45 ttl=128 time=2.20 ms
64 bytes from 10.0.0.65: icmp_seq=46 ttl=128 time=2.14 ms
64 bytes from 10.0.0.65: icmp_seq=47 ttl=128 time=2.09 ms
64 bytes from 10.0.0.65: icmp_seq=48 ttl=128 time=2.50 ms
64 bytes from 10.0.0.65: icmp_seq=49 ttl=128 time=4.07 ms
64 bytes from 10.0.0.65: icmp_seq=50 ttl=128 time=1.86 ms
64 bytes from 10.0.0.65: icmp_seq=51 ttl=128 time=2.73 ms
64 bytes from 10.0.0.65: icmp_seq=52 ttl=128 time=6.45 ms
64 bytes from 10.0.0.65: icmp_seq=53 ttl=128 time=2.33 ms
64 bytes from 10.0.0.65: icmp_seq=54 ttl=128 time=1.89 ms
64 bytes from 10.0.0.65: icmp_seq=55 ttl=128 time=1.56 ms
64 bytes from 10.0.0.65: icmp_seq=56 ttl=128 time=11.6 ms
64 bytes from 10.0.0.65: icmp_seq=57 ttl=128 time=6.92 ms
64 bytes from 10.0.0.65: icmp_seq=58 ttl=128 time=2.28 ms
64 bytes from 10.0.0.65: icmp_seq=59 ttl=128 time=4.05 ms
64 bytes from 10.0.0.65: icmp_seq=60 ttl=128 time=2.58 ms
64 bytes from 10.0.0.65: icmp_seq=61 ttl=128 time=4.98 ms
64 bytes from 10.0.0.65: icmp_seq=62 ttl=128 time=1.92 ms

This is not in modem sleep at all.

Has this ever worked properly?

I do not see any bug fixes related to modem sleep for 3.0.4 and 3.0.5.

This is a legacy design, updating to the RTOS SDK is not currently an option.

Is there some other operating condition I need to meet in order for the sleep mode to take effect? Has it ever worked? Are there any examples that show it working?