Any plans to update to Kernel 5.10.x?

[fernandodiacenco]

Hi there,

FQ_CODEL_FAST compiles normally on a Debian 10 router Kernel 4.19, but it does not compile on the same Debian running 5.10 Kernel, giving the error:

make: Entering directory '/usr/src/linux-headers-5.10.0-0.bpo.5-amd64'
CC [M] /usr/src/fq_codel_fast/sch_fq_codel_fast.o
In file included from /usr/src/fq_codel_fast/sch_fq_codel_fast.c:26:
/usr/src/linux-headers-5.10.0-0.bpo.5-common/include/net/pkt_cls.h:913:22: error: ‘TCQ_ETS_MAX_BANDS’ undeclared here (not in a function); did you mean ‘TCQ_MIN_PRIO_BANDS’?
unsigned int quanta[TCQ_ETS_MAX_BANDS]; /* 0 for strict bands. */
^~~~~~~~~~~~~~~~~
TCQ_MIN_PRIO_BANDS
/usr/src/fq_codel_fast/sch_fq_codel_fast.c: In function ‘fq_codel_enqueue’:
/usr/src/fq_codel_fast/sch_fq_codel_fast.c:162:2: warning: parameter names (without types) in function declaration
int uninitialized_var(ret);
^~~
/usr/src/fq_codel_fast/sch_fq_codel_fast.c:168:37: error: ‘ret’ undeclared (first use in this function); did you mean ‘net’?
idx = fq_codel_classify(skb, sch, &ret);
^~~
net
/usr/src/fq_codel_fast/sch_fq_codel_fast.c:168:37: note: each undeclared identifier is reported only once for each function it appears in
make[2]: *** [/usr/src/linux-headers-5.10.0-0.bpo.5-common/scripts/Makefile.build:284: /usr/src/fq_codel_fast/sch_fq_codel_fast.o] Error 1
make[1]: *** [/usr/src/linux-headers-5.10.0-0.bpo.5-common/Makefile:1819: /usr/src/fq_codel_fast] Error 2
make: *** [/usr/src/linux-headers-5.10.0-0.bpo.5-common/Makefile:185: __sub-make] Error 2
make: Leaving directory '/usr/src/linux-headers-5.10.0-0.bpo.5-amd64'

Any plans to update to 5.10? I ask because I only managed to get fq_pie on kernel 5.10 and I kinda want to test fq_codel_fast along with fq_pie

Thanks.

[dtaht]

put a

#define TCQ_ETS_MAX_BANDS 16

in somewhere before it hits this call.

yes I do hope to resume testing fq_codel_fast at some point... and it would be great to know if it still worked at all on a later OS. the GRO splitting has become more and more important of late.

[fernandodiacenco]

Now it almost compiles, but still complains about a variable called ret

sudo make -C "/usr/src/linux-headers-uname -r" M=/usr/src/fq_codel_fast modules

make: Entering directory '/usr/src/linux-headers-5.10.0-0.bpo.7-amd64'
CC [M] /usr/src/fq_codel_fast/sch_fq_codel_fast.o
/usr/src/fq_codel_fast/sch_fq_codel_fast.c: In function ‘fq_codel_enqueue’:
/usr/src/fq_codel_fast/sch_fq_codel_fast.c:162:2: warning: parameter names (without types) in function declaration
int uninitialized_var(ret);
^~~
/usr/src/fq_codel_fast/sch_fq_codel_fast.c:168:37: error: ‘ret’ undeclared (first use in this function); did you mean ‘net’?
idx = fq_codel_classify(skb, sch, &ret);
^~~
net
/usr/src/fq_codel_fast/sch_fq_codel_fast.c:168:37: note: each undeclared identifier is reported only once for each function it appears in
make[2]: *** [/usr/src/linux-headers-5.10.0-0.bpo.7-common/scripts/Makefile.build:284: /usr/src/fq_codel_fast/sch_fq_codel_fast.o] Error 1
make[1]: *** [/usr/src/linux-headers-5.10.0-0.bpo.7-common/Makefile:1845: /usr/src/fq_codel_fast] Error 2
make: *** [/usr/src/linux-headers-5.10.0-0.bpo.7-common/Makefile:185: __sub-make] Error 2
make: Leaving directory '/usr/src/linux-headers-5.10.0-0.bpo.7-amd64'

Looking at the file sch_fq_codel_fast.c I see a variable called ret, so its not a typo as suggested by the error code, I don't know what to do.

[dtaht]

I am trying to clear time to work on this more fully. However I would be very interested in your test plan and test results for fq_pie, fq_codel, and cake, in the interim?

round 'here we use flent a lot. and sqm-scripts to configure itr

[fernandodiacenco]

I wrote a big text explaining everything, but realized I need to hone the skill of getting a point across without too much text

Want to flent rrul, my flent broken, not fixed yet
Using iperf3 and dslreports and gut feeling to test
Home connection 100/50, starts with about 95mbps upload then averages to 50mbps in 10sec or so
Can I set a limit that takes account of this variation? Trying HTB and TBF with ceil/burst, not sure if doing right*
How they work without setting upper limit: cake good, fq_codel good, fq_pie bad
How they work setting a upper limit: all good
Gut is liking fq_codel, less resources spent than cake
Gut has crush on fq_codel_fast, want to date, invite broken

TLDR: I'm trying to understand the algorithms behavior to see if its feasible to set up a bunch of used wireless access points (with Openwrt or anything else, no wireless) as a sort of SQM appliance, to improve my community/peers quality of experience as they are working from home while the kids watches classes online and stuff.

*This is what I'm doing on the limiters

With TBF
tc qdisc add dev ppp0 root handle 1: tbf rate 50mbit burst 90mbit latency 300ms
tc qdisc add dev ppp0 parent 1: handle 110: fq_codel

<Repeat for fq_pie and cake, but cake is: cake dual-dsthost overhead 44>

With HTB
tc qdisc add dev ppp0 root handle 1: htb default 1
tc class add dev ppp0 parent 1: classid 1:1 htb rate 90mbit ceil 90mbit
tc class add dev ppp0 parent 1:1 classid 1:10 htb rate 50mbit ceil 90mbit
tc qdisc add dev ppp0 parent 1:10 fq_codel

[dtaht]

see the sqm-scripts