ath: is here full list of supported chipsets and chipsets comparsion?

Sam Leffler sam at freebsd.org
Mon Dec 15 15:01:43 PST 2008


Wes Morgan wrote:
> On Mon, 15 Dec 2008, Sam Leffler wrote:
>
>> Lev Serebryakov wrote:
>>> Hello, All.
>>>
>>>   `man ath' on FreeBSD 7.1-PRE speaks only about WPA2 in AR5212 and
>>> not-supported AR5005VL. But in "current" cars here are many other
>>> chipsets -- 5213A, 5414, etc... And Atheros site is not very helpful
>>> now  -- there are not 5212, 5213A, 5414 chipsets in both areas "WLAN
>>> for Home, Office and Metro Wi-Fi" and "WLAN for Mobile" (BTW, link to
>>> http://customerproducts.atheros.com/ doesn't work anymore).
>>>
>>>   Is here full list of supported chipsets, and, maybe, some table with
>>> chipsets features (AES, WPA2, AP mode, etc)?
>>>
>>>
>> HEAD supports most PCI/cardbus parts.  The main exceptions are the 
>> 9280 and 9285.  The ath9k driver for linux supports them and anyone 
>> can add support using that.  11n parts only support legacy operation 
>> though w/ ~10 line change to the driver you can get 11n RX + legacy TX.
>>
>> RELENG_7 has a much older hal and lacks support for many cards.  I 
>> recommend using HEAD if wireless support is important to you.  No ETA 
>> on an update by me--others are welcome to supply the changes.
>>
>> All ath cards support all features you listed (except for the 5210 
>> which you're unlikely to care about).
>
> Sam,
>
> I just updated my system from -stable to -current this weekend, and 
> I'm noticing a lot more issues with the ath driver losing its 
> association much more frequently, sometimes failing to reassociate 
> altogether. The strange part is that tcpdump shows packets being 
> received from the network just fine, but nothing seems to be 
> transmitted (although I have not stepped over to my file server to 
> verify this). A manual unload / reload of the module "solves" the 
> problem, but I get a warning about a memory leak.

I see no relevant PR's.  I cannot fix problems w/o information.

>
> Code-wise, the only difference since the "open sourcing" is simply 
> that we now have the code, correct?

Only difference between what?  The code in the tree is my latest work.  
If there are problems I will do my best to fix them given sufficient 
information and/or the ability to reproduce the problem.

    Sam



More information about the freebsd-net mailing list