[PATCH] mesh: Skip STA expiration on user_mpm mode

[PATCH] mesh: Skip STA expiration on user_mpm mode

[Masashi Honma]

The mesh STA expiration processing works even if user_mpm mode. This causes
unexpected STA disconnection.

This patch skips STA expiration processing on user_mpm mode.

Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
---
 net/mac80211/mesh.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/mac80211/mesh.c b/net/mac80211/mesh.c
index 0c8b2a7..50174fd 100644
--- a/net/mac80211/mesh.c
+++ b/net/mac80211/mesh.c
@@ -574,7 +574,8 @@ static void ieee80211_mesh_housekeeping(struct ieee80211_sub_if_data *sdata)
 	struct ieee80211_if_mesh *ifmsh = &sdata->u.mesh;
 	u32 changed;
 
-	ieee80211_sta_expire(sdata, ifmsh->mshcfg.plink_timeout * HZ);
+	if (!ifmsh->user_mpm)
+		ieee80211_sta_expire(sdata, ifmsh->mshcfg.plink_timeout * HZ);
 	mesh_path_expire(sdata);
 
 	changed = mesh_accept_plinks_update(sdata);
-- 
2.1.0

Re: [PATCH] mesh: Skip STA expiration on user_mpm mode

[Bob Copeland]

On Tue, Jan 13, 2015 at 9:23 PM, Masashi Honma <masashi.honma@gmail.com> wrote:
> The mesh STA expiration processing works even if user_mpm mode. This causes
> unexpected STA disconnection.
>
> This patch skips STA expiration processing on user_mpm mode.
>
> Signed-off-by: Masashi Honma <masashi.honma@gmail.com>

>
> -       ieee80211_sta_expire(sdata, ifmsh->mshcfg.plink_timeout * HZ);
> +       if (!ifmsh->user_mpm)
> +               ieee80211_sta_expire(sdata, ifmsh->mshcfg.plink_timeout * HZ);

I took a look at what authsae is doing, and it looks like it's just
prepared to handle the NL80211_CMD_DEL_STATION events generated from
cfg80211_del_sta() -- is there a reason that couldn't work for wpa_s
too?  Then you'd only need the patch to configure that timeout on the
client, and you can simplify it by not caring about user_mpm.

-- 
Bob Copeland %% www.bobcopeland.com

Re: [PATCH] mesh: Skip STA expiration on user_mpm mode

[Masashi Honma]

2015-01-15 12:14 GMT+09:00 Bob Copeland <me@bobcopeland.com>:
> I took a look at what authsae is doing, and it looks like it's just
> prepared to handle the NL80211_CMD_DEL_STATION events generated from
> cfg80211_del_sta() -- is there a reason that couldn't work for wpa_s
> too?  Then you'd only need the patch to configure that timeout on the
> client, and you can simplify it by not caring about user_mpm.

Thank you for your review.

Indeed wpa_supplicant do nothing even if it receivced NL80211_CMD_DEL_STATION.
But the mismatch (wpa_supplicant recognizes STA is alive, mac80211
recognizes STA leaved) is not so good.

Anyway I have not considered authsae user land application.
Does authsae have it's own inactivity timer ?
If not, I will drop this patch.

Re: [PATCH] mesh: Skip STA expiration on user_mpm mode

[Bob Copeland]

On Fri, Jan 16, 2015 at 11:04:30AM +0900, Masashi Honma wrote:
> 2015-01-15 12:14 GMT+09:00 Bob Copeland <me@bobcopeland.com>:
> > I took a look at what authsae is doing, and it looks like it's just
> > prepared to handle the NL80211_CMD_DEL_STATION events generated from
> > cfg80211_del_sta() -- is there a reason that couldn't work for wpa_s
> > too?  Then you'd only need the patch to configure that timeout on the
> > client, and you can simplify it by not caring about user_mpm.
> 
> Thank you for your review.
> 
> Indeed wpa_supplicant do nothing even if it receivced NL80211_CMD_DEL_STATION.
> But the mismatch (wpa_supplicant recognizes STA is alive, mac80211
> recognizes STA leaved) is not so good.

Right - so wpa_s could be changed to handle this event (for mesh),
I think.

> Anyway I have not considered authsae user land application.
> Does authsae have it's own inactivity timer ?
> If not, I will drop this patch.

No, it relies on mac80211's inactivity timer.  So, yeah, I think this
patch would break authsae, in the sense that stations would never
get removed.

-- 
Bob Copeland %% http://bobcopeland.com/

Re: [PATCH] mesh: Skip STA expiration on user_mpm mode

[Masashi Honma]

2015-01-16 21:59 GMT+09:00 Bob Copeland <me@bobcopeland.com>:
> No, it relies on mac80211's inactivity timer.  So, yeah, I think this
> patch would break authsae, in the sense that stations would never
> get removed.

I see. I will drop this patch.