On Wed, 2008-11-19 at 16:38 -0800, Abhijeet Kolekar wrote:
> Patch fixes the kernel trace when user tries to set
> ad-hoc mode on non IBSS channel.
> e.g iwconfig wlan0 chan 36 mode ad-hoc
>
> Signed-off-by: Abhijeet Kolekar <abhijeet.kolekar@xxxxxxxxx>
> ---
> net/mac80211/wext.c | 10 ++++++++++
> 1 files changed, 10 insertions(+), 0 deletions(-)
>
> diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c
> index 63f36e9..5a9bc8c 100644
> --- a/net/mac80211/wext.c
> +++ b/net/mac80211/wext.c
> @@ -271,6 +271,7 @@ static int ieee80211_ioctl_siwmode(struct net_device *dev,
> __u32 *mode, char *extra)
> {
> struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
> + struct ieee80211_local *local = sdata->local;
> int type;
>
> if (sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
> @@ -281,6 +282,15 @@ static int ieee80211_ioctl_siwmode(struct net_device *dev,
> type = NL80211_IFTYPE_STATION;
> break;
> case IW_MODE_ADHOC:
> + /* Following check added to avoid misconfiguration if
> + * you set mode and frequency at the same time.
> + */
The way that comment is written it belongs in the changelog.
> + if (local->oper_channel && local->oper_channel->flags &
> + IEEE80211_CHAN_NO_IBSS) {
> + printk(KERN_DEBUG "Not a valid IBSS channel\n");
And that's some extremely odd indentation/wrapping there. Also, imho the
printk is unnecessary.
johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
