>did work with onoe.
>
>Thanks for your continued effort
>
>
>//JT
>
>-----Ursprungligt meddelande-----
>Fran: John Bicket [mailto:jbicket@amsterdam.lcs.mit.edu]
>Skickat: den 10 mars 2005 06:48
>Till: Jonas Tarnstrom
>Kopia: Bruno Randolf;
madwifi-devel@lists.sourceforge.net
>Amne: Re: SV: [Madwifi-devel] rate control algorithm
>
>
>
>Hi guys,
>
>I setup an access point and found the bug. I wasn't resetting the
>management rates correctly in ap mode. Attached is a patch that fixes
>that, and moves the ath_rate_setmgtrates into the ath_rate_ctl_reset
>function.
>
>--jbicket
>
>
>Jonas Tarnstrom [jonas.tarnstrom@repeatit.se] wrote:
>>I did some further investigation after all (it hard to leave the office
>>tonight).
>>
>>Having the client running ath_rate_sample and the AP running ath_rate_onoe
>>works but when I switch to sample on both it stops to work, this seems to
>be
>>an AP issue.
>>
>>//JT
>>
>>-----Ursprungligt meddelande-----
>>Fran:
madwifi-devel-admin@lists.sourceforge.net
>>[mailto:madwifi-devel-admin@lists.sourceforge.net]For Jonas Tarnstrom
>>Skickat: den 9 mars 2005 17:55
>>Till: John Bicket
>>Kopia: Bruno Randolf;
madwifi-devel@lists.sourceforge.net
>>Amne: RE: [Madwifi-devel] rate control algorithm
>>
>>
>>Hi,
>>Bad news, while the association now works no data frames gets through.
>Tried
>>a simple ping.
>>My setup is (roughly):
>>
>>AP:
>>Madwifi (modified CVS, dating back to Oct 2004), 802.11a channel 36. Rate
>>AUTO
>>
>>Client:
>>Madwifi (same as above)
>>
>>Will investigate this further tomorrow.
>>Though my code tree differs from the BSD branch ONOE and AMRR works
>>perfectly in the same setup.
>>
>>//JT
>>
>>
>>-----Ursprungligt meddelande-----
>>Fran: John Bicket [mailto:jbicket@amsterdam.lcs.mit.edu]
>>Skickat: den 9 mars 2005 16:51
>>Till: Jonas Tarnstrom
>>Kopia: Bruno Randolf;
madwifi-devel@lists.sourceforge.net
>>Amne: Re: [Madwifi-devel] rate control algorithm
>>
>>
>>Hi,
>>
>>Here is a patch that has bunch of fixes:
>>
>>1. use ath_hal_computetxtime so that it will estimate throughput for
>>turbo modes correctly.
>>2. calculate the txtime independently for a few different packet sizes
>>since delivery probability is affected by packet size.
>>3. fix setting the bit-rate for management frames.
>>4. fix ewma to average the first few packets instead of just setting the
>>average to the first packet.
>>
>>I have tested this with an 802.11b, 802.11a, and 802.11g access point;
>>if you still have problems can you tell me your set-up so I can
>>replicate it?
>>
>>Thanks,
>>
>>--john
>>
>>
>>Jonas Tarnstrom [jonas.tarnstrom@repeatit.se] wrote:
>>>Hi,
>>>I've applied your patch but my two madwifi clients still won't associate
>to
>>>each other.
>>>
>>>Is there work in progress regarding this?
>>>
>>>//JT
>>>
>>>
>>>-----Ursprungligt meddelande-----
>>>Fran:
madwifi-devel-admin@lists.sourceforge.net
>>>[mailto:madwifi-devel-admin@lists.sourceforge.net]For John Bicket
>>>Skickat: den 7 mars 2005 18:33
>>>Till: Bruno Randolf
>>>Kopia:
madwifi-devel@lists.sourceforge.net
>>>Amne: Re: [Madwifi-devel] rate control algorithm
>>>
>>>
>>>
>>>Hi Bruno,
>>>Here is a patch that fixes the scanning problem, and also updates
>>>ni_txrate when it switches bitrates.
>>>
>>>Thanks,
>>>
>>>John
>>>
>>>
>>>
>>>
>>>Bruno Randolf [bruno.randolf@4g-systems.biz] wrote:
>>>>hello john!
>>>>
>>>>the sample rate control seems to work only in B mode in my setup. in A
>and
>>>G
>>>>mode the client cannot associate with the AP and scans forever.
>>>>
>>>>nevertheless, since it's optional and does not break anything, i have
>>added
>>>>the patch to the CVS HEAD and BSD branch. on the BSD branch i had to
>>change
>>>>one little thing, to adapt for the new ath_softc struct:
>>>>
>>>>- int rate = sc->sc_hwmap[ds->ds_txstat.ts_rate &
>>>IEEE80211_RATE_VAL];
>>>>+ int rate = sc->sc_hwmap[ds->ds_txstat.ts_rate &
>>>>IEEE80211_RATE_VAL].ieeerate;
>>>>
>>>>greetings,
>>>>bruno
>>>>
>>>>On Sunday 06 March 2005 22:21, John Bicket wrote:
>>>>> Hi Stefan,
>>>>>
>>>>> I just put a pdf at
http://www.pdos.lcs.mit.edu/papers/jbicket-ms.pdf
>>>>> using the -Ppdf switch.
>>>>>
>>>>> I also will keep an updated version of the patch at:
>>>>>
http://www.pdos.lcs.mit.edu/~jbicket/sample_rate.patch
>>>>>
>>>>> Thanks,
>>>>>
>>>>> John
>>>>>
>>>>> Stefan Br??ns [lurchi@kawo1.rwth-aachen.de] wrote:
>>>>> >Am Sonntag, 6. M??rz 2005 07:15 schrieb John Bicket:
>>>>> >> [...]
>>>>> >> If you're interested in the details, my thesis on this
>>>>> >> (including lots of measurements) is at
>>>>> >>
http://www.pdos.lcs.mit.edu/papers/jbicket-ms.ps
>>>>> >
>>>>> >Nice to read ...
>>>>> >But can you create a pdf version of this? Don't forget the "-Ppdf"
>>>switch
>>>>> > for dvips, else it will use bitmap versions of the fonts and not the
>>>>> > type1 (vector). After this, you can use ghostscript for the ps2pdf
>>>>> > conversion.
>>>>> >
>>>>> >> I hope people will find this useful. I'm sure there are bugs in the
>>>>> >> code that I haven't experienced yet, so if you see any, please let
>me
>>>>> >> know.
>>>>> >>
>>>>> >> Thanks!
>>>>> >>
>>>>> >> John Bicket
>>>>> >
>>>>> >Greetings,
>>>>> >
>>>>> >Stefan
>>>>> >
>>>>> >--
>>>>> >Stefan Br??ns / Kastanienweg 6 - Zimmer 1206 / 52074 Aachen
>>>>> >mailto:lurch@gmx.li
http://www.kawo1.rwth-aachen.de/~lurchi/
>>>>> > phone: +49 241 169-4206 mobile: +49 160 3797725
>>>>>
>>>>> -------------------------------------------------------
>>>>> SF email is sponsored by - The IT Product Guide
>>>>> Read honest & candid reviews on hundreds of IT Products from real
>users.
>>>>> Discover which products truly live up to the hype. Start reading now.
>>>>>
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>>>>> _______________________________________________
>>>>> Madwifi-devel mailing list
>>>>>
Madwifi-devel@lists.sourceforge.net
>>>>>
https://lists.sourceforge.net/lists/listinfo/madwifi-devel
>>>
>>>
>>>
>>>
>>>
>>>-------------------------------------------------------
>>>SF email is sponsored by - The IT Product Guide
>>>Read honest & candid reviews on hundreds of IT Products from real users.
>>>Discover which products truly live up to the hype. Start reading now.
>>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>>>_______________________________________________
>>>Madwifi-devel mailing list
>>>Madwifi-devel@lists.sourceforge.net
>>>https://lists.sourceforge.net/lists/listinfo/madwifi-devel
>>
>>
>>
>>-------------------------------------------------------
>>SF email is sponsored by - The IT Product Guide
>>Read honest & candid reviews on hundreds of IT Products from real users.
>>Discover which products truly live up to the hype. Start reading now.
>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>>_______________________________________________
>>Madwifi-devel mailing list
>>Madwifi-devel@lists.sourceforge.net
>>https://lists.sourceforge.net/lists/listinfo/madwifi-devel
>>
>>
>>
>>--
>>No virus found in this incoming message.
>>Checked by AVG Anti-Virus.
>>Version: 7.0.308 / Virus Database: 266.7.0 - Release Date: 2005-03-08
>>
>>
>>
>>
>>-------------------------------------------------------
>>SF email is sponsored by - The IT Product Guide
>>Read honest & candid reviews on hundreds of IT Products from real users.
>>Discover which products truly live up to the hype. Start reading now.
>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>>_______________________________________________
>>Madwifi-devel mailing list
>>Madwifi-devel@lists.sourceforge.net
>>https://lists.sourceforge.net/lists/listinfo/madwifi-devel
>
>
>
>-------------------------------------------------------
>SF email is sponsored by - The IT Product Guide
>Read honest & candid reviews on hundreds of IT Products from real users.
>Discover which products truly live up to the hype. Start reading now.
>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>_______________________________________________
>Madwifi-devel mailing list
>Madwifi-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/madwifi-devel
