Why should we implement/use our watchdog instead of Linux default watchdog?

Question

I am studying Android driver/kernel. I find that there are 2 set of watchdogs:

1. kernel/watchdog.c
2. /drivers/watchdog/tegra_wdt.c

Both them are used to recovery the system as system hang on an event (eg infinite loop).

The first one is enabled in system boot time by default, but many people still implementing their watchdog in /drivers/watchdog/ folder for specific driver. I confuse about this. Why should we need multiple watchdog? Is any one know about the "philosophy" about this?

Answer 1

因为看门狗用于安全线程，所以借助此功能，我们可以停止线程，使其不再能够穿孔。