Testing this now

Appears to work correctly:

Test Branch Changes Merged into Dev cleanly. This is ready for next release.

@dave14305 feel free to merge this at your connivence; you may want to push an update on SNB forums; or not. Depending how you feel.

I don’t think these changes warrant a new release yet.

I don’t think these changes warrant a new release yet.

How about a way to warn users installing on 3006 with BE class routers that Adaptive QoS is currently unfunctional?

Do you have some key indicators to identify if adaptive QoS is "broken" or not?
We could use those and make the warning adaptive in that case! Or just put in a temporary warning based on the router firmware and class.

Do you have some key indicators to identify if adaptive QoS is "broken" or not?

Listing and counting the tc filter show dev br0 output is the most obvious method to me.

Listing and counting the tc filter show dev br0 output is the most obvious method to me.

This is already in the script. It’s one of the tests I use to determine if Adaptive QoS is done loading all its rules before FlexQoS tries to modify them.

Listing and counting the tc filter show dev br0 output is the most obvious method to me.

This is already in the script. It’s one of the tests I use to determine if Adaptive QoS is done loading all its rules before FlexQoS tries to modify them.

FlexQoS/flexqos.sh

Line 1592 in d6ae4c2

check_qos_tc() {

I'm thinking we piggy back on that then, and then maybe make the warning check dependent on 3006 firmware being installed. Or I guess technically we could make it a universal warning no matter the firmware installed, this could be an issue in the future maybe even if it gets fixed eventually?

Or I guess technically we could make it a universal warning no matter the firmware installed, this could be an issue in the future maybe even if it gets fixed eventually

I expect that most BE owners that try to run it are getting the messages from this check after the timeout.

Or I guess technically we could make it a universal warning no matter the firmware installed, this could be an issue in the future maybe even if it gets fixed eventually

I expect that most BE owners that try to run it are getting the messages from this check after the timeout.

FlexQoS/flexqos.sh

Lines 1698 to 1712 in d6ae4c2

	while check_qos_tc;
	do
	[ "${sleepdelay}" = "0" ] && logmsg "TC Modification Delayed Start"
	sleep 10s
	if [ "${sleepdelay}" -ge "180" ]; then
	logmsg "QoS state: Classes=${dlclasscnt} | Filters=${dlfiltercnt} | qdiscs=${dlqdisccnt}"
	if [ ! -f "/tmp/${SCRIPTNAME}_restartonce" ]; then
	touch "/tmp/${SCRIPTNAME}_restartonce"
	logmsg "TC Modification Delay reached maximum 180 seconds. Restarting QoS."
	service "restart_qos;restart_firewall"
	else
	logmsg "TC Modification Delay reached maximum 180 seconds again. Canceling startup!"
	rm "/tmp/${SCRIPTNAME}_restartonce" 2>/dev/null
	fi
	return 1

I would just put a big splash warning on the main menu for the users to let them know myself. The logs is a good place to validate though I can see if I get those messages

I expect that most BE owners that try to run it are getting the messages from this check after the timeout.

FlexQoS/flexqos.sh

Lines 1698 to 1712 in d6ae4c2

	while check_qos_tc;
	do
	[ "${sleepdelay}" = "0" ] && logmsg "TC Modification Delayed Start"
	sleep 10s
	if [ "${sleepdelay}" -ge "180" ]; then
	logmsg "QoS state: Classes=${dlclasscnt} | Filters=${dlfiltercnt} | qdiscs=${dlqdisccnt}"
	if [ ! -f "/tmp/${SCRIPTNAME}_restartonce" ]; then
	touch "/tmp/${SCRIPTNAME}_restartonce"
	logmsg "TC Modification Delay reached maximum 180 seconds. Restarting QoS."
	service "restart_qos;restart_firewall"
	else
	logmsg "TC Modification Delay reached maximum 180 seconds again. Canceling startup!"
	rm "/tmp/${SCRIPTNAME}_restartonce" 2>/dev/null
	fi
	return 1

The logs is a good place to validate though I can see if I get those messages

@dave14305

Confirmed as seen below:

Jun  7 14:47:01 FlexQoS: TC Modification Delay reached maximum 180 seconds. Restarting QoS.
Jun  7 14:47:01 rc_service: service 6760:notify_rc restart_qos;restart_firewall
Jun  7 14:47:01 custom_script: Running /jffs/scripts/service-event (args: restart qos)
Jun  7 14:47:02 BWDPI: fun bitmap = 3
Jun  7 14:47:02 custom_script: Running /jffs/scripts/service-event-end (args: restart qos)
Jun  7 14:47:02 custom_script: Running /jffs/scripts/service-event (args: restart firewall)
Jun  7 14:47:04 rc_service: waitting "restart_qos;restart_firewall"(last_rc:restart_qos;restart_firewall) via  ...
Jun  7 14:47:05 custom_script: Running /jffs/scripts/firewall-start (args: ppp0)
Jun  7 14:47:05 custom_script: Running /jffs/scripts/service-event-end (args: restart firewall)
Jun  7 14:47:05 FlexQoS: /jffs/addons/flexqos/flexqos.sh (pid=8184) called in unattended mode with 1 args: -start
Jun  7 14:47:05 FlexQoS: Applying iptables static rules
Jun  7 14:47:05 FlexQoS: Applying iptables custom rules
Jun  7 14:47:06 FlexQoS: Flushing conntrack table
Jun  7 14:47:06 FlexQoS: TC Modification Delayed Start

@dave14305

Quick daft/sample here: #8

Closing this since I’ve probably made a mess of this one by not understanding the intended workflow.

Closing this since I’ve probably made a mess of this one by not understanding the intended workflow.

It's not that bad truly. As soon as I merged mine into dev it was included in this PR.
We can open a new one against master though when enough changes make it worth while.

Thanks for your help. Hope all works well with OPNSense