After doing
...but the device still responds to pings and TCP connect request via ethernet. If no serial console is attached to see the final message, it's impossible to know if it has safely shut down and power can be disconnected.
Seems odd that the kernel seems to be still running, I guess so it can detect a keypress and initiate a reboot - but why not just have a small section of dedicated code for this? Why would ethernet interrupts and TCP/IP packets still need to be handled?
Curious whether this is a Raspberry Pi thing, or it's more generic and happens with all systems that cannot be powered off by the OS. (Perhaps most tier 1 hardware supporting that may explain why there's no dedicated "shut everything down except serial and/or keyboard interrupts" code.)
I notice the power LED switches off during boot. Would be handy if this could stay on, and instead was switched off at the point FreeBSD prints the final halt message to the console?
shutdown -p now
or shutdown -h now
, serial console says:
Code:
The operating system has halted.
Please press any key to reboot.
...but the device still responds to pings and TCP connect request via ethernet. If no serial console is attached to see the final message, it's impossible to know if it has safely shut down and power can be disconnected.
Seems odd that the kernel seems to be still running, I guess so it can detect a keypress and initiate a reboot - but why not just have a small section of dedicated code for this? Why would ethernet interrupts and TCP/IP packets still need to be handled?
Curious whether this is a Raspberry Pi thing, or it's more generic and happens with all systems that cannot be powered off by the OS. (Perhaps most tier 1 hardware supporting that may explain why there's no dedicated "shut everything down except serial and/or keyboard interrupts" code.)
I notice the power LED switches off during boot. Would be handy if this could stay on, and instead was switched off at the point FreeBSD prints the final halt message to the console?