linux-wireless.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* iw station del broken?
@ 2010-03-29 12:47 Marco Porsch
  2010-03-29 12:58 ` Johannes Berg
  0 siblings, 1 reply; 6+ messages in thread
From: Marco Porsch @ 2010-03-29 12:47 UTC (permalink / raw)
  To: linux-wireless

Hi list,

is it just me or is 'iw dev <devname> station del <MAC address>' broken?

I am using it in mesh mode to build generic topologies out of a fully 
meshed net. I am using it first since Javiers latest patch "[PATCH v2] 
mac80211: Handle mesh action frames in ieee80211_rx_h_action" which 
fixed mesh mode successfully.

I didn't try it in anything other than mesh. Running latest OpenWrt with 
compat-wireless-2010-03-24 and kernel 2.6.32.10 including Javiers patch. 
Maybe a fix for nl80211 is missing?


root@OpenWrt:/# iw dev imesh5g station get xx:xx:xx:xx:xx:xx
Station xx:xx:xx:xx:xx:xx (on imesh5g)
        inactive time:  410 ms
        rx bytes:       8569
        rx packets:     132
        tx bytes:       321
        tx packets:     5
        signal:         -19 dBm
        tx bitrate:     6.0 MBit/s
        mesh llid:      2784
        mesh plid:      64165
        mesh plink:     ESTAB
root@OpenWrt:/# iw dev imesh5g station del xx:xx:xx:xx:xx:xx
command failed: Invalid argument (-22)


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: iw station del broken?
  2010-03-29 12:47 iw station del broken? Marco Porsch
@ 2010-03-29 12:58 ` Johannes Berg
  2010-03-29 13:05   ` Marco Porsch
  0 siblings, 1 reply; 6+ messages in thread
From: Johannes Berg @ 2010-03-29 12:58 UTC (permalink / raw)
  To: Marco Porsch; +Cc: linux-wireless

On Mon, 2010-03-29 at 14:47 +0200, Marco Porsch wrote:
> Hi list,
> 
> is it just me or is 'iw dev <devname> station del <MAC address>' broken?

No ... we decided a while ago that station manipulation in mesh was too
hard to support and removed it.

What's the use case anyway?

johannes


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: iw station del broken?
  2010-03-29 12:58 ` Johannes Berg
@ 2010-03-29 13:05   ` Marco Porsch
  2010-03-29 13:09     ` Johannes Berg
  0 siblings, 1 reply; 6+ messages in thread
From: Marco Porsch @ 2010-03-29 13:05 UTC (permalink / raw)
  To: Johannes Berg; +Cc: linux-wireless

Johannes Berg schrieb:
> On Mon, 2010-03-29 at 14:47 +0200, Marco Porsch wrote:
>   
>> Hi list,
>>
>> is it just me or is 'iw dev <devname> station del <MAC address>' broken?
>>     
>
> No ... we decided a while ago that station manipulation in mesh was too
> hard to support and removed it.
>
> What's the use case anyway?
>
> johannes
>   
The use case is that, if I have some mesh points in a single room or one 
a small floor, I can simply give them a topology by removing the links 
between the STA, I do not wish to be directly connected.
Removing the links means 'iw dev <devname> set mesh_param 
mesh_auto_open_plinks 0', which blocks building up new peering links, 
and then removing some links with 'station del'.

Hmpf.. I used this feature quiet often...

Regards,
Marco

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: iw station del broken?
  2010-03-29 13:05   ` Marco Porsch
@ 2010-03-29 13:09     ` Johannes Berg
  2010-03-29 13:41       ` Marco Porsch
  0 siblings, 1 reply; 6+ messages in thread
From: Johannes Berg @ 2010-03-29 13:09 UTC (permalink / raw)
  To: Marco Porsch; +Cc: linux-wireless

On Mon, 2010-03-29 at 15:05 +0200, Marco Porsch wrote:
> Johannes Berg schrieb:
> > On Mon, 2010-03-29 at 14:47 +0200, Marco Porsch wrote:
> >   
> >> Hi list,
> >>
> >> is it just me or is 'iw dev <devname> station del <MAC address>' broken?
> >>     
> >
> > No ... we decided a while ago that station manipulation in mesh was too
> > hard to support and removed it.
> >
> > What's the use case anyway?
> >
> > johannes
> >   
> The use case is that, if I have some mesh points in a single room or one 
> a small floor, I can simply give them a topology by removing the links 
> between the STA, I do not wish to be directly connected.
> Removing the links means 'iw dev <devname> set mesh_param 
> mesh_auto_open_plinks 0', which blocks building up new peering links, 
> and then removing some links with 'station del'.

Can't you just do that before it gets discovered?

I'm not really happy with the mesh API anyway ...

johannes


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: iw station del broken?
  2010-03-29 13:09     ` Johannes Berg
@ 2010-03-29 13:41       ` Marco Porsch
  2010-03-29 13:43         ` Johannes Berg
  0 siblings, 1 reply; 6+ messages in thread
From: Marco Porsch @ 2010-03-29 13:41 UTC (permalink / raw)
  To: Johannes Berg; +Cc: linux-wireless

Johannes Berg schrieb:
> On Mon, 2010-03-29 at 15:05 +0200, Marco Porsch wrote:
>   
>> Johannes Berg schrieb:
>>     
>>> On Mon, 2010-03-29 at 14:47 +0200, Marco Porsch wrote:
>>>   
>>>       
>>>> Hi list,
>>>>
>>>> is it just me or is 'iw dev <devname> station del <MAC address>' broken?
>>>>     
>>>>         
>>> No ... we decided a while ago that station manipulation in mesh was too
>>> hard to support and removed it.
>>>
>>> What's the use case anyway?
>>>
>>> johannes
>>>   
>>>       
>> The use case is that, if I have some mesh points in a single room or one 
>> a small floor, I can simply give them a topology by removing the links 
>> between the STA, I do not wish to be directly connected.
>> Removing the links means 'iw dev <devname> set mesh_param 
>> mesh_auto_open_plinks 0', which blocks building up new peering links, 
>> and then removing some links with 'station del'.
>>     
>
> Can't you just do that before it gets discovered?
>
> I'm not really happy with the mesh API anyway ...
>
> johannes
>
>   
That makes any scripting impossible, because it would mean a fine timing 
between the ifup on the stations. Also I think you can't build complex 
topologies this way.
The alternative iw dev <devname> mpath set <destination MAC address> 
next_hop <next hop MAC address> is not useful (if it is still included 
though ^^) because it sets paths hard. Means that these paths are not 
changed, even if they are broken.

Well, you're the maintainer. Your choice.

By the way: which was the last version with 'station del' in mesh 
intact? Which codeparts would have to be exchanged if i'd like to use 
this further? Only iw and nl80211?

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: iw station del broken?
  2010-03-29 13:41       ` Marco Porsch
@ 2010-03-29 13:43         ` Johannes Berg
  0 siblings, 0 replies; 6+ messages in thread
From: Johannes Berg @ 2010-03-29 13:43 UTC (permalink / raw)
  To: Marco Porsch; +Cc: linux-wireless

On Mon, 2010-03-29 at 15:41 +0200, Marco Porsch wrote:

> That makes any scripting impossible, because it would mean a fine timing 
> between the ifup on the stations. Also I think you can't build complex 
> topologies this way.

Hmm, you're probably right.

> The alternative iw dev <devname> mpath set <destination MAC address> 
> next_hop <next hop MAC address> is not useful (if it is still included 
> though ^^) because it sets paths hard. Means that these paths are not 
> changed, even if they are broken.
> 
> Well, you're the maintainer. Your choice.
> 
> By the way: which was the last version with 'station del' in mesh 
> intact? Which codeparts would have to be exchanged if i'd like to use 
> this further? Only iw and nl80211?

Actually I think you only need to change a line in nl80211.c allowing
deleting; in fact deleting is OK (and thus I think we can put a patch in
adding it back) -- adding was the problem.

johannes


^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2010-03-29 13:43 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-03-29 12:47 iw station del broken? Marco Porsch
2010-03-29 12:58 ` Johannes Berg
2010-03-29 13:05   ` Marco Porsch
2010-03-29 13:09     ` Johannes Berg
2010-03-29 13:41       ` Marco Porsch
2010-03-29 13:43         ` Johannes Berg

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).