________________________________
> Date: Tue, 22 Apr 2008 18:26:34 +0200
> From: [EMAIL PROTECTED]
> To: [EMAIL PROTECTED]
> Subject: Re: Feedback on ASUS WL-138G V2
> CC: [EMAIL PROTECTED]; [email protected]; [EMAIL PROTECTED]
>
>
>
> On Tue, Apr 22, 2008 at 5:24 PM, Johannes Berg <[EMAIL PROTECTED]> wrote:
>
>> Notice the lack of led devices registering here (as I mentioned before).
>> Indeed,
>> the behaviour of the leds is totally different -- now, they come ON and stay
>> ON when the card is 'active'...ie; ifconfig wlan0 up. Converse is true ;
>> doing
>> 'ifconfig wlan0 down' turns the leds off again.
>>
>> If I dump the sprom image Stefanik provided back into the card, the led
>> device
>> registration lines reappear in debug, and the leds blink in that sympathetic
>> way
>> rx/tx leds do. Is there yet another funky value in sprom responsible for
>> this?
>
> Yes, the sprom determines how the card vendor wants the LEDs to behave.
> You can change this behaviour easily in sysfs.
>
> And as we can expect from ASUS, the LED settings are wrong, too. (All are @
> 0xFF.) The same is true for the antenna bitfields (it has 1 antenna for B/G,
> and no A antennas, yet the bitfield says 0 B/G antennas and 2 A ones...).
> ASUS has done a terrible work with the SPROM (or perhaps these are the
> default values of the SPROM that Broadcom programmed into the chips before
> they reached ASUS - who only changed the subsys IDs). The workaround is to
> always enable the LEDs for ASUS cards, while the real fix is up to ASUS - it
> would be firing the current engineer (who doesn't seem to know the word
> "SPROM" and connects Bluetooth GPIO lines to 802.11 TX amplifiers - I imagine
> if he was a surgeon, we would have a lot of people who died because he
> connected their liver where their heart would have gone...) and getting a
> more competent one to do the work. :-)
>
....Haha, sad fact is that surgeons like that actually exist, and end up
killing dozens of people
before they loose their job - remember, I'm in .AU and we've seen a spate of
this sort of
thing over the past few years -- fortunately, I don't think the asus cards
present any health
risk....
....It does make one consider the insanity of all this though....ie;
given the state of
'programming finesse' demonstrated by asus with these cards, one
doesn't>really< have
to change b43 code at all (to account for asus' misgivings) -- all we need do
is correctly
identify this particular type of card and change sprom contents to something
that works....
I know that's not a 'sane' comment -- I'm just considering how much (good) work
the
developers here do (and -will- do) to the b43 code to get the driver working
correctly
with these cards -- the b43 code will increase in size by so many kilobytes of
work-
arounds, fixups and extra routines, all to circumvent a lousy 258bytes of sprom
code
which itself is about as useful as having square wheels on a car....
>
>> [from the 'host' box with the asus card]
>>
>> 1212 packets transmitted, 1200 packets received, 0% packet loss
>> round-trip min/avg/max/stddev = 0.821/1.192/5.801/0.366 ms
>>
>> ...???.....I would've expected almost 1% packet loss (having lost 12 packets)
>
> Heh. May be a rounding error somewhere
>
> Apparently it always rounds down, and doesn't follow the 5/4 rule.
.....oh, I was unaware of this....
>
>> If I ping the host box from my machine here, the packets have to go out
>> my eth0 interface into a netgear wr602 in client mode which is associated
>> with the wgr614 AP down the hall (which is the AP the host is associated
>> with also). In this case, I get ;
>
>> 350 packets transmitted, 344 packets received, +140 duplicates, 1% packet
>> loss
>> round-trip min/avg/max/stddev = 1.222/1.893/4.296/0.443 ms
>>
>> Has anyone any idea what's going on here?
>
> Eh, what exactly is your network setup? You normally can't bridge STA
> mode. This looks like something is not doing ieee 802.11 de-duplication.
>
> Donald, did you have a radiotap interface on the same card while you did
> this? I tend to get some dupe pings when I run kismet in the background (with
> channel hopping>disabled, of course).
No....no radiotap, kismet or anything else running like that at the time.
Regards,
Donald
_________________________________________________________________
It's simple! Sell your car for just $30 at CarPoint.com.au
http://a.ninemsn.com.au/b.aspx?URL=http%3A%2F%2Fsecure%2Dau%2Eimrworldwide%2Ecom%2Fcgi%2Dbin%2Fa%2Fci%5F450304%2Fet%5F2%2Fcg%5F801459%2Fpi%5F1004813%2Fai%5F859641&_t=762955845&_r=tig_OCT07&_m=EXT
_______________________________________________
Bcm43xx-dev mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
