Jörg Sommer wrote:
> Hi,
>
> the code in bcm43xx_wc.c:bcm43xx_wx_set_channelfreq() passes the value
> unchanged to bcm43xx_freq_to_channel() which leads to mistakes, because
> bcm43xx_freq_to_channel() expects the frequency in MHz, but the value
> given to bcm43xx_wx_set_channelfreq() might be scaled somehow different.
>
> Also passing the internal frequency value in bcm43xx_wx_get_rangeparams()
> to userland is not correct. The frequency is in MHz, but the factor
> passed to userland is freq*10^1.
>
> #v+
> --- /home/joerg/bcm43xx_wx.c 2007-02-07 11:25:19.000000000 +0100
> +++ bcm43xx_wx.c 2007-02-07 11:09:46.000000000 +0100
> @@ -108,16 +108,21 @@
> u8 channel;
> int freq;
> int err = -EINVAL;
> + s16 ex;
>
> mutex_lock(&bcm->mutex);
> spin_lock_irqsave(&bcm->irq_lock, flags);
>
> - if ((data->freq.m >= 0) && (data->freq.m <= 1000)) {
> + if ((data->freq.e == 0) &&
> + (data->freq.m >= 0) && (data->freq.m <= 1000)) {
> channel = data->freq.m;
> freq = bcm43xx_channel_to_freq(bcm, channel);
> } else {
> - channel = bcm43xx_freq_to_channel(bcm, data->freq.m);
> freq = data->freq.m;
> + ex = 6 - data->freq.e;
> + while (--ex >= 0) /* scale down the frequency to MHz */
> + freq /= 10;
> + channel = bcm43xx_freq_to_channel(bcm, freq);
> }
> if (!ieee80211_is_valid_channel(bcm->ieee, channel))
> goto out_unlock;
I'm testing this hunk now.
> @@ -295,7 +300,7 @@
> break;
> range->freq[j].i = j + 1;
> range->freq[j].m = geo->bg[i].freq;//FIXME?
> - range->freq[j].e = 1;
> + range->freq[j].e = 6;
> j++;
> }
> range->num_frequency = j;
This one has already been fixed by multiplying geo->bg[i].freq by 100,000. Same
result. There was a
similar error for 802.11a frequencies as well.
> @@ -439,7 +444,7 @@
> u16 maxpower;
>
> if ((data->txpower.flags & IW_TXPOW_TYPE) != IW_TXPOW_DBM) {
> - printk(PFX KERN_ERR "TX power not in dBm.\n");
> + printk(KERN_ERR PFX "TX power not in dBm.\n");
> return -EOPNOTSUPP;
> }
Good catch. Thanks for your contribution.
Larry
_______________________________________________
Bcm43xx-dev mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
