Please Test: Updated Ral Driver Patch for rt2860/rt3090

[PseudoCylon]

> ------------------------------
>
> Message: 12
> Date: Sun, 11 Mar 2012 14:14:23 -0700
> From: Adrian Chadd <adrian at freebsd.org>
> Subject: Re: Please Test: Updated Ral Driver Patch for rt2860/rt3090
> To: "Sevan / Venture37" <venture37 at gmail.com>
> Cc: freebsd-wireless at freebsd.org
> Message-ID:
>        <CAJ-Vmo=jJmwY=bTVaKFYTd--1E9vzwgn54VMJyY82bc9pSWcBg at mail.gmail.com>
> Content-Type: text/plain; charset=ISO-8859-1
>
> On 11 March 2012 14:10, Sevan / Venture37 <venture37 at gmail.com> wrote:
>
>>> I'll post a new one in a bit. If you want to help, would you be ok with
>>> git as opposed to hg?
>>
>>
>> nope not at all but if its because of familiarity then in this scenario it
>> doesn't make much difference which is used, same work flow, change, commit,
>> push, pull, rinse, repeat. :)
>
> Hiya,
>
> Let's just strip the ^M's from the tree and as long as ray@ and
> pseudocyclon are a-ok with the work thus far, I'll commit it to -HEAD.
>
> We can then tidy things up from there.
>
> I just don't want to commit ^M's to the tree, or blatant style(9)
> violations. But we can always tidy things up incrementally once
> they're in the tree.
>

I've talked about ieee80211_send_bar() with matt, already.

Other than that,
* rt2860_raw_xmit() ignores ieee80211_bpf_params,
* wouldn't that be better to disable some interrupts that only
increment counters. i.e rt2860_gp_timer_intr()

Though, we can tidy up all of those later.


> The driver attaches but ifconfig wlan0 scan doesn't show any access
> points

All mgt frames are sent as wme class of 5, but probe_req frames should
be sent as WME_AC_BE. This might cause glitch in probing. Hence
nothing in scan result. (?)


AK