debugging 802.11s

[Andrey Yurovsky]

A quick update on this: commit
dc300b5c5291d779a53d83af960fdf9388bf0544 seems to have broken ath5k
mesh, so anything up till that commit (and after
d059b2aed9e93cb94364e01374bca90ce3dc0ef9) should be working.

  -Andrey

On Wed, Nov 12, 2008 at 2:09 PM, Andrey Yurovsky <andrey at cozybit.com> wrote:
> Ok, sounds great.  I believe we then have a regression somewhere after
> that commit and HEAD which breaks ath5k mesh however.
>
>  -Andrey
>
> On Wed, Nov 12, 2008 at 1:21 PM, Keir <keirlawson at gmail.com> wrote:
>> Yes, the mesh appears to be working (IE I managed to ping from each node to
>> the other), I am using commit d059b2aed9e93cb94364e01374bca90ce3dc0ef9 as
>> you suggested.
>>
>> Cheers
>>
>> Keir
>>
>> 2008/11/12 Andrey Yurovsky <andrey at cozybit.com>
>>>
>>> Hi Keir, no problem.  So just to be clear:
>>> 1) does mesh work for you?
>>> 2) what git commit are you using?
>>>
>>> Thanks!
>>>
>>>  -Andrey
>>>
>>> On Wed, Nov 12, 2008 at 8:25 AM, Keir <keirlawson at gmail.com> wrote:
>>> > Apologies, I've made a fool of myself, my firewall was blocking the
>>> > pings
>>> > all along :-P *embarrassed*
>>> >
>>> > Sorry for waisting your time
>>> >
>>> > Keir
>>> >
>>> > 2008/11/12 Keir <keirlawson at gmail.com>
>>> >>
>>> >> Still no luck, still getting the sendmsg permission denied error on one
>>> >> box, the other box ping is just timing out.  How does o11s play with
>>> >> 11g or
>>> >> b networks? could that be the problem?  arp shows that it's managed to
>>> >> map
>>> >> IPs to MACs correctly, when I do a station dump both stations now have
>>> >> plink
>>> >> status of ESTAB.  I'm going to start sticking printks about the mesh
>>> >> code,
>>> >> unless there is something more useful I can do?
>>> >>
>>> >> Cheers
>>> >>
>>> >> Keir
>>> >>
>>> >> 2008/11/11 Andrey Yurovsky <andrey at cozybit.com>
>>> >>>
>>> >>> d059b2aed9e93cb94364e01374bca90ce3dc0ef9 should work and I would
>>> >>> suspect that it's broken some time after.  Can you try that one?
>>> >>>
>>> >>> On Tue, Nov 11, 2008 at 1:26 PM, Keir <keirlawson at gmail.com> wrote:
>>> >>> > Erm, if I knew what git-bisect was ;-) (still havent got git fully
>>> >>> > worked
>>> >>> > out) what commit would you recommend I try?
>>> >>> >
>>> >>> > Cheers
>>> >>> >
>>> >>> > Keir
>>> >>> >
>>> >>> > 2008/11/11 Andrey Yurovsky <andrey at cozybit.com>
>>> >>> >>
>>> >>> >> Hi Keir.  I just rebuilt the latest wireless-testing and ath5k
>>> >>> >> isn't
>>> >>> >> working for me either (in fact I don't even see beacons on the
>>> >>> >> air).
>>> >>> >> Perhaps a recent regression?  Do you have time to git-bisect this
>>> >>> >> or
>>> >>> >> at least try an earlier commit?  Thanks,
>>> >>> >>
>>> >>> >>  -Andrey
>>> >>> >>
>>> >>> >> On Tue, Nov 11, 2008 at 8:58 AM, Keir <keirlawson at gmail.com> wrote:
>>> >>> >> > I also spotted this in dmesg:
>>> >>> >> >
>>> >>> >> > [ 3603.303963] ath5k phy0: beacon queue 7 didn't stop?
>>> >>> >> > [ 3605.761677] ath5k phy0: beacon queue 7 didn't stop?
>>> >>> >> > [ 3656.554337] ath5k phy0: beacon queue 7 didn't stop?
>>> >>> >> > [ 3657.280923] Inbound IN=eth1 OUT=
>>> >>> >> > MAC=00:90:27:43:08:2a:00:10:b5:e8:04:d3:08:00 SRC=192.168.0.251
>>> >>> >> > DST=192.168.0.1 LEN=328 TOS=0x00 PREC=0x00 TTL=64 ID=0 DF
>>> >>> >> > PROTO=UDP
>>> >>> >> > SPT=68
>>> >>> >> > DPT=67 LEN=308
>>> >>> >> > [ 3763.209246] Unknown OutputIN= OUT=mesh0 SRC=10.0.0.1
>>> >>> >> > DST=10.0.0.2
>>> >>> >> > LEN=84
>>> >>> >> > TOS=0x00 PREC=0x00 TTL=64 ID=0 DF PROTO=ICMP TYPE=8 CODE=0
>>> >>> >> > ID=44569
>>> >>> >> > SEQ=1
>>> >>> >> > [ 3764.218103] Unknown OutputIN= OUT=mesh0 SRC=10.0.0.1
>>> >>> >> > DST=10.0.0.2
>>> >>> >> > LEN=84
>>> >>> >> > TOS=0x00 PREC=0x00 TTL=64 ID=0 DF PROTO=ICMP TYPE=8 CODE=0
>>> >>> >> > ID=44569
>>> >>> >> > SEQ=2
>>> >>> >> >
>>> >>> >> > Just after attempting to ping from one to the other.
>>> >>> >> >
>>> >>> >> > Keir
>>> >>> >> >
>>> >>> >> > 2008/11/11 Keir <keirlawson at gmail.com>
>>> >>> >> >>
>>> >>> >> >>> Hi Keir.  With NetworkManager not running, can you try things
>>> >>> >> >>> again
>>> >>> >> >>> from the top?
>>> >>> >> >>>
>>> >>> >> >>> 1) reload the driver
>>> >>> >> >>> # modprobe -r ath5k
>>> >>> >> >>> # modprobe ath5k
>>> >>> >> >>>
>>> >>> >> >>> 2) bring up your mesh interfaces, ex:
>>> >>> >> >>> # iw dev wlan0 interface add mesh0 type mp mesh_id mymesh
>>> >>> >> >>> # ifconfig mesh0 up
>>> >>> >> >>> # ifconfig mesh0 10.0.0.1
>>> >>> >> >>> (etc)
>>> >>> >> >>>
>>> >>> >> >>> ...and then ping again?
>>> >>> >> >>> # ping -c 2 -I mesh0 10.0.0.2
>>> >>> >> >>
>>> >>> >> >> I get the same error as before :-(
>>> >>> >> >>>
>>> >>> >> >>> Also, what's the git tag for your kernel (see Luis'
>>> >>> >> >>> instructions
>>> >>> >> >>> on
>>> >>> >> >>> how to obtain this via git-describe).  Thanks,
>>> >>> >> >>
>>> >>> >> >>
>>> >>> >> >> keir at dave:~/Desktop/Code/wireless-testing$ git describe
>>> >>> >> >> v2.6.28-rc2-4151-g43288a0
>>> >>> >> >> keir at dave:~/Desktop/Code/wireless-testing$ git tag -l| grep
>>> >>> >> >> master
>>> >>> >> >> |
>>> >>> >> >> tail
>>> >>> >> >> -1
>>> >>> >> >> master-2008-10-27
>>> >>> >> >>
>>> >>> >> >> station dump looks like:
>>> >>> >> >> root at dave:/home/keir/Desktop/Code/iw-0.9.6# ./iw dev mesh0
>>> >>> >> >> station
>>> >>> >> >> dump
>>> >>> >> >> Station 00:18:4d:7a:51:70 (on mesh0)
>>> >>> >> >>     inactive time:    320 ms
>>> >>> >> >>     rx bytes:    27301
>>> >>> >> >>     tx bytes:    13844
>>> >>> >> >>     mesh llid:    1378
>>> >>> >> >>     mesh plid:    30482
>>> >>> >> >>     mesh plink:    HOLDING
>>> >>> >> >>
>>> >>> >> >> Any ideas?
>>> >>> >> >>
>>> >>> >> >> Cheers
>>> >>> >> >>
>>> >>> >> >> Keir
>>> >>> >> >
>>> >>> >
>>> >>> >
>>> >>
>>> >
>>> >
>>
>



-- 
Andrey Yurovsky
cozybit Inc.