HFSC QoS script sets uplink wrong

[phuzi0n]

The HFSC script configures both interfaces with tc classes that are the speed entered for your downlink. The code in svqos.hfsc shows that it's using $2 to set both interfaces. Also, there's inconsistent usage of $1 and $2 in svqos.htb and svqos.hfsc. It seems that svqos.htb has the downlink set in $1 but svqos.hfsc has the downlink set in $2 which leads to more confusion.

http://www.dd-wrt.com/phpBB2/viewtopic.php?t=52339

For users wishing to correct this themselves, configure the GUI QoS however you like but leave it set to use HTB. Then go to the Administration->Commands page and save this startup and firewall script. If you want you can remove the startup script and store it somewhere permanent such as JFFS.

#Startup
cd /tmp
wget http://phuzi0n.googlecode.com/svn/trunk/svqos.hfsc
chmod +x svqos.hfsc

#Firewall
if [ 1 == `nvram get wshaper_enable` ]; then
/tmp/svqos.hfsc `nvram get wshaper_uplink` `nvram get wshaper_downlink` `nvram get wan_iface` `nvram get wan_mtu` 0
fi

[BrainSlayer]

dont try to analyse these scripts if you dont know how they are working

[phuzi0n]

I guess that's why you just corrected one of the mistakes then, because I just don't understand it? The HFSC script has major problems and if you didn't read then link then here's the evidence plain and simple.

Here's my classes with up 384kbps / down 1536 kbps.

root@DD-WRT:~# tc class show dev imq0
class hfsc 1: root
class hfsc 1:1 parent 1: sc m1 0bit d 0us m2 100000Kbit ul m1 0bit d 0us m2 100000Kbit
class hfsc 1:100 parent 1:1 leaf 100: ls m1 100000Kbit d 100.0ms m2 100000Kbit ul m1 0bit d 0us m2 100000Kbit
class hfsc 1:10 parent 1:1 leaf 10: rt m1 0bit d 6.1ms m2 307000bit ls m1 0bit d 9.0ms m2 1152Kbit ul m1 0bit d 0us m2 1536Kbit
class hfsc 1:20 parent 1:1 leaf 20: rt m1 0bit d 7.5ms m2 460000bit ls m1 0bit d 9.6ms m2 1152Kbit ul m1 0bit d 0us m2 1536Kbit
class hfsc 1:30 parent 1:1 leaf 30: rt m1 324328bit d 37.0ms m2 153000bit ls m1 0bit d 21.4ms m2 768000bit ul m1 0bit d 0us m2 1536Kbit
class hfsc 1:40 parent 1:1 leaf 40: ls m1 153000bit d 500.0ms m2 153000bit ul m1 0bit d 0us m2 1382Kbit

root@DD-WRT:~# tc class show dev eth1
class hfsc 1: root
class hfsc 1:1 parent 1: sc m1 0bit d 0us m2 100000Kbit ul m1 0bit d 0us m2 100000Kbit
class hfsc 1:100 parent 1:1 leaf 100: ls m1 100000Kbit d 100.0ms m2 100000Kbit ul m1 0bit d 0us m2 100000Kbit
class hfsc 1:10 parent 1:1 leaf 10: rt m1 0bit d 6.1ms m2 307000bit ls m1 0bit d 9.0ms m2 1152Kbit ul m1 0bit d 0us m2 1536Kbit
class hfsc 1:20 parent 1:1 leaf 20: rt m1 0bit d 7.5ms m2 460000bit ls m1 0bit d 9.6ms m2 1152Kbit ul m1 0bit d 0us m2 1536Kbit
class hfsc 1:30 parent 1:1 leaf 30: rt m1 324328bit d 37.0ms m2 153000bit ls m1 0bit d 21.4ms m2 768000bit ul m1 0bit d 0us m2 1536Kbit
class hfsc 1:40 parent 1:1 leaf 40: ls m1 153000bit d 500.0ms m2 153000bit ul m1 0bit d 0us m2 1382Kbit

[phuzi0n]

Line 637 and 639 clearly show that ul_val and dl_val are being passed in a different order to each script.

​http://svn.dd-wrt.com:8000/dd-wrt/browser/src/router/services/networking/wshaper.c?rev=11938#L637

[phuzi0n]

Here's a corrected HFSC script.

​http://phuzi0n.googlecode.com/svn/trunk/svqos.hfsc

[jstein]

I have the same problem as phuzi0n. I fixed it with a custom script to correct for the ppp0 interface. \

[phuzi0n]

Please don't ignore this bug.

[anonymous]

Milestone v24 final deleted

[phuzi0n]

Fixed in r14390