And it turns out the current head has this fixed already...
I guess this must be Kevin's fix, I could have sworn that I had that in my
repository when I started testing, but I guess I tried juggling a few
repositories too many on too many computers.
So it seems this was fixed all along, well it made me try git bisect so all was
good...
Just for documentation:
1) Building (from inside the sch_cake repository:
sudo tc qdisc delete root dev eth0 ; sudo rmmod sch_cake ; tc -s qdisc show dev
eth0 ; make ; sudo make install
2) Testing:
sudo ../iproute2-cake-next/tc/tc qdisc delete root dev eth0 ; sudo
../iproute2-cake-next/tc/tc qdisc add root dev eth0 cake raw overhead 0 ;
../iproute2-cake-next/tc/tc -s qdisc show dev eth0
Results
user@work-horse:~/CODE/sch_cake$ sudo ../iproute2-cake-next/tc/tc qdisc delete
root dev eth0 ; sudo ../iproute2-cake-next/tc/tc qdisc add root dev eth0 cake
raw overhead 0 ; ../iproute2-cake-next/tc/tc -s qdisc show dev eth0
[...]
GOOD: qdisc cake 8035: root refcnt 6 unlimited diffserv3 triple-isolate rtt
100.0ms raw total_overhead 14 hard_header_len 14
BAD: qdisc cake 8035: root refcnt 6 unlimited diffserv3 triple-isolate rtt
100.0ms raw total_overhead 14 hard_header_len 14 # note with the added
"raw" keyword this always worked
[...]
user@work-horse:~/CODE/sch_cake$ sudo ../iproute2-cake-next/tc/tc qdisc delete
root dev eth0 ; sudo ../iproute2-cake-next/tc/tc qdisc add root dev eth0 cake
overhead 0 ; ../iproute2-cake-next/tc/tc -s qdisc show dev eth0
[...]
GOOD: qdisc cake 8036: root refcnt 6 unlimited diffserv3 triple-isolate rtt
100.0ms noatm overhead 0 via-ethernet total_overhead 0 hard_header_len 14
BAD: qdisc cake 8036: root refcnt 6 unlimited diffserv3 triple-isolate rtt
100.0ms noatm overhead 14 via-ethernet total_overhead 14 hard_header_len 14
[...]
user@work-horse:~/CODE/sch_cake$ sudo ../iproute2-cake-next/tc/tc qdisc delete
root dev eth0 ; sudo ../iproute2-cake-next/tc/tc qdisc add root dev eth0 cake
ethernet ; ../iproute2-cake-next/tc/tc -s qdisc show dev eth0
[...]
GOOD: qdisc cake 8037: root refcnt 6 unlimited diffserv3 triple-isolate rtt
100.0ms noatm overhead 38 via-ethernet total_overhead 38 hard_header_len 14 mpu
84
BAD: qdisc cake 8037: root refcnt 6 unlimited diffserv3 triple-isolate rtt
100.0ms noatm overhead 14 via-ethernet total_overhead 14 hard_header_len 14 mpu
84
[...]
The BAD results are from memory so might be off in the "overhead N" and
"via-ethernet" results but the total_overhead and hard_header_len looked quite
similar to the reported ones...
Best Regards
Sebastian
> On Nov 23, 2017, at 23:02, Sebastian Moeller <[email protected]> wrote:
>
> And git bisect gave:
>
> user@work-horse:~/CODE/sch_cake$ git bisect bad
> 031998e4eee58cbc706711eba8c54684f07306be is the first bad commit
> commit 031998e4eee58cbc706711eba8c54684f07306be
> Author: Dave Taht <[email protected]>
> Date: Sun Nov 19 19:02:06 2017 -0800
>
> sch_cake: make more checkpatch compliant
>
> :100644 100644 97a4727e6ed9b9451567568b027e65dd9937112c
> e6a4b09ae8bebab5ae07b94b469184669ab6e27f M sch_cake.c
> user@work-horse:~/CODE/sch_cake$
>
> as the likely culprit, now let's see whether I can figure out why. (The git
> bisect thing, even though I had to skip a few commits makes bug hunting
> almost fun).
>
>
> Best Regards
>
>> On Nov 23, 2017, at 22:43, Sebastian Moeller <[email protected]> wrote:
>>
>> 9789742cfc596d48583ba4cdbc8f38d026121fa6
>
> _______________________________________________
> Cake mailing list
> [email protected]
> https://lists.bufferbloat.net/listinfo/cake
_______________________________________________
Cake mailing list
[email protected]
https://lists.bufferbloat.net/listinfo/cake
