> On 3 Mar 2019, at 12:22, John Sager <[email protected]> wrote: > > If you are going to do that, I would suggest using a few of the upper bits > of the 32-bit fwmark/connmark space available, rather than the lowest bits. > Then that would allow to use fwmarks other purposes, and to use the lowest > bits, as well in the future. As iptables allows a mask before comparison, > then choose a specific mask for the bits you use both for setting and testing.
That’s a good point and I’m sort of hoping upstream reject the current submission on that basis. I think the ‘use of fwmarks’ needs more thought as to how it’s done for the future - too keen to get something out. Maybe it’s sufficient as is, I don’t know. What I do know is that after implementing the ‘bad idea’ I subsequently implemented the very bad idea! Using the top byte of mark, bits 5-0 are the DSCP, bit 6 is the ‘Cake set this’ flag, and bit 7 is left alone as it tends to get mis-interpreted as a sign bit. I implemented the iptables marking chains as ‘check if cake fwmark bit set, go to a DSCP mangle chain if not’, mangle the DSCP, picked up by CAKE which (in egress mode) copies the DSCP to the fwmark’ and set the ‘cake did this’ bit. In theory connections only go through the dscp mangle once. Cake in ingress mode will copy the fwmark if it’s been set by cake back into the diffserv field on each packet. I think that last step should be made optional but I’ve had enough ‘fun’ for the day. https://github.com/ldir-EDB0/sch_cake/commits/mine Cheers, Kevin D-B 012C ACB2 28C6 C53E 9775 9123 B3A2 389B 9DE2 334A _______________________________________________ Cake mailing list [email protected] https://lists.bufferbloat.net/listinfo/cake
