[rt2x00-users] [RFC] rt2x00: rt2800: enable STBC

Ivo Van Doorn ivdoorn at gmail.com
Fri May 28 23:27:06 AEST 2010


On Fri, May 28, 2010 at 3:25 PM, Ivo Van Doorn <ivdoorn at gmail.com> wrote:
> On Fri, May 28, 2010 at 3:18 PM, Helmut Schaa
> <helmut.schaa at googlemail.com> wrote:
>> Am Freitag 28 Mai 2010 schrieb Ivo Van Doorn:
>>> On Fri, May 28, 2010 at 3:09 PM, Helmut Schaa
>>> <helmut.schaa at googlemail.com> wrote:
>>> > Am Freitag 28 Mai 2010 schrieb Ivo Van Doorn:
>>> >> On Fri, May 28, 2010 at 2:31 PM, Helmut Schaa
>>> >> <helmut.schaa at googlemail.com> wrote:
>>> >> > Am Freitag 28 Mai 2010 schrieb Gertjan van Wingerde:
>>> >> >> On 05/28/10 13:35, Helmut Schaa wrote:
>>> >> >> > +   if (tx_info->flags & IEEE80211_TX_CTL_STBC)
>>> >> >> > +           __set_bit(ENTRY_TXD_HT_STBC, &txdesc->flags);
>>> >> >> >
>>> >> >> >     /*
>>> >> >> >      * Determine HT Mix/Greenfield rate mode
>>> >> >>
>>> >> >> This doesn't seem to work. In mac80211 the IEEE80211_TX_CTL_STBC is actually a 2-bit field,
>>> >> >> so is the field in the TXWI structure of rt2x00. I suggest we keep the stbc field a u16 (or a u8)
>>> >> >> and simply copy the value of these two bits into that field.
>>> >> >
>>> >> > I know what you mean but the ralink drivers only use 0 and 1. Maybe because
>>> >> > there are no 3 tx stream devices yet? But you're right, let's keep the stbc
>>> >> > field and map everything !=0 to 1 for now?
>>> >>
>>> >> You might want to check my STBC patch in the experimental branch of rt2x00.git,
>>> >> I had coded it yesterday and does use the entire IEEE80211_TX_CTL_STBC field
>>> >> of the flag (Note that IEEE80211_TX_CTL_STBC is also 2 bits).
>>> >
>>> > Funny ;)
>>> >
>>> > Nevertheless, which branch do you mean, the one on git.kernel.org? I didn't
>>> > find the commit.
>>>
>>> It was apparently the commit I made _after_ I pushed the rebase of
>>> rt2x00.git to the server.
>>
>> :)
>>
>>> Wait for a few minutes, and the patch will appear, since I have pushed
>>> it out now.
>>
>> Ok, thanks, got it.
>>
>> I guess, we'll just take it as you've done it. Looks much cleaner.
>>
>> Nevertheless I'm not sure what the device will do with values other then 1? At
>> least the vendor driver only uses 0 and 1.
>
> Well it depends on the number of streams right, and don't we tell mac80211
> the number of supported streams during initialization? If it turns out that the
> third stream is not supported, we just have to fix that in the initialization.

P.S. Since you already agree on this patch I'll post the patch to
rt2x00-users later
today so it can be merged into the main branch. At the moment the experimental
tree is there for my private work only, and will not be merged into
master without
review on rt2x00-users.

Ivo




More information about the users mailing list