It seems that at some point in the last three months (between r37989 and r39097), something about the way network interfaces are brought up changed. Interfaces that were formerly available at a certain point during startup are no longer necessarily available at that point.
I first noticed this when I found that dnsmasq wasn’t responding to DNS queries. I have dnsmasq configured to bind to each desired interface’s address individually rather than to bind to to the wildcard address. This is accomplished by setting “option nonwildcard 1” in /etc/config/dhcp’s “config dnsmasq” section. With this setting in place, after boot, dnsmasq was not bound to port 53 on any address. I checked this with “netstat -an”. This is affecting some of my other startup scripts as well, which had been expecting the system to be in a certain state with respect to network interface availability based on their START= values. It seems that network devices are now brought up in the background, and may not even be available when any rc.d start script runs. Is this a new change with the ongoing netifd or procd work? In the new procd/netifd world, is there a better way to start services that depend on specific network interfaces?
_______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel