search

beckdac / ESP8266-transparent-bridge

Absolutely transparent bridge for the ESP8266
GNU General Public License v2.0
367 stars 95 forks source link

strange statement causing latency #14

Closed tve closed 9 years ago

tve commented 9 years ago

Can you explain what the following line does:

WRITE_PERI_REG(0X60000914, 0x73); //WTD

from https://github.com/beckdac/ESP8266-transparent-bridge/blob/master/user/user_main.c#L39

It is necessary as far as I can tell, but there is something weird happening overall which causes a lot of latency and I can't find any description or mention of register 0X60000914 anywhere.

ArnieO commented 9 years ago

According to the brief comment on the link below, it seems to be linked to the watchdog timer? https://github.com/willemwouters/ESP8266/blob/master/sdk/esp_iot_sdk_v0.9.3/examples/at_v0.20_on_SDKv0.9.3/at/user/at_port.c

tve commented 9 years ago

Ah, thanks I hadn't seen that version of the AT example. The next version, e.g. https://github.com/espressif/esp8266_at/blob/master/at/driver/uart.c doesn't have the WDT line in it anymore. The best I could find about the actual WDT registers is https://github.com/themadinventor/esp-blobless/blob/master/include/esp8266.h which indicates that this line resets the WDT. Why that would be needed here is beyond me, but what I observe is that the whole RX code construction introduces 200-300ms latency. Something's off...

tve commented 9 years ago

I've partially resolved my issue, it's not the WDT or uart that is causing the latency but the sent callback that is very late. Now trying to find out why...