| Message ID | 1359853341-29237-3-git-send-email-thomas@cozybit.com (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
On Sat, 2013-02-02 at 17:02 -0800, Thomas Pedersen wrote: > - pos = skb_put(skb, ie_ssid_len); > - *pos++ = WLAN_EID_SSID; > + if (ieee80211_vif_is_mesh(vif)) { > + pos = skb_put(skb, 2 + 2 + ssid_len); > + *pos++ = WLAN_EID_SSID; > + *pos++ = 0; > + /* NOTE: mesh ID will be out of order */ Why put it out of order? Also I'm not convinced that it's a good idea to translate "SSID" from the userspace API to "mesh ID" silently? Might make more sense to have those separately maybe? I mean, it seems reasonable to even think you might scan for a mesh network when you're not a mesh interface, for example? johannes -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Feb 4, 2013 at 9:38 AM, Johannes Berg <johannes@sipsolutions.net> wrote: > On Sat, 2013-02-02 at 17:02 -0800, Thomas Pedersen wrote: > >> - pos = skb_put(skb, ie_ssid_len); >> - *pos++ = WLAN_EID_SSID; >> + if (ieee80211_vif_is_mesh(vif)) { >> + pos = skb_put(skb, 2 + 2 + ssid_len); >> + *pos++ = WLAN_EID_SSID; >> + *pos++ = 0; >> + /* NOTE: mesh ID will be out of order */ > > Why put it out of order? > > Also I'm not convinced that it's a good idea to translate "SSID" from > the userspace API to "mesh ID" silently? Might make more sense to have > those separately maybe? I mean, it seems reasonable to even think you > might scan for a mesh network when you're not a mesh interface, for > example? Yeah that makes more sense, but will obviously require more work. I'll just drop this for now.
On Mon, Feb 4, 2013 at 9:38 AM, Johannes Berg <johannes@sipsolutions.net> wrote: > On Sat, 2013-02-02 at 17:02 -0800, Thomas Pedersen wrote: > >> - pos = skb_put(skb, ie_ssid_len); >> - *pos++ = WLAN_EID_SSID; >> + if (ieee80211_vif_is_mesh(vif)) { >> + pos = skb_put(skb, 2 + 2 + ssid_len); >> + *pos++ = WLAN_EID_SSID; >> + *pos++ = 0; >> + /* NOTE: mesh ID will be out of order */ > > Why put it out of order? > > Also I'm not convinced that it's a good idea to translate "SSID" from > the userspace API to "mesh ID" silently? Might make more sense to have > those separately maybe? I mean, it seems reasonable to even think you > might scan for a mesh network when you're not a mesh interface, for > example? Yeah that makes more sense, but will obviously require more work. I'll just drop this for now.
On Mon, 2013-02-04 at 09:58 -0800, Thomas Pedersen wrote: > On Mon, Feb 4, 2013 at 9:38 AM, Johannes Berg <johannes@sipsolutions.net> wrote: > > On Sat, 2013-02-02 at 17:02 -0800, Thomas Pedersen wrote: > > > >> - pos = skb_put(skb, ie_ssid_len); > >> - *pos++ = WLAN_EID_SSID; > >> + if (ieee80211_vif_is_mesh(vif)) { > >> + pos = skb_put(skb, 2 + 2 + ssid_len); > >> + *pos++ = WLAN_EID_SSID; > >> + *pos++ = 0; > >> + /* NOTE: mesh ID will be out of order */ > > > > Why put it out of order? > > > > Also I'm not convinced that it's a good idea to translate "SSID" from > > the userspace API to "mesh ID" silently? Might make more sense to have > > those separately maybe? I mean, it seems reasonable to even think you > > might scan for a mesh network when you're not a mesh interface, for > > example? > > Yeah that makes more sense, but will obviously require more work. > I'll just drop this for now. I guess the other question is if you actually want this at all. I mean, if you just do patch 1 and 3, then unless you want to actively scan for multiple networks you can just include the mesh ID in the IE parameter. Seems for many purposes that could be acceptable already. Now I'm not saying that you shouldn't put it into the kernel, there certainly could be value in that, just saying that in terms of effort that might be easier? johannes -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Feb 4, 2013 at 10:03 AM, Johannes Berg <johannes@sipsolutions.net> wrote: > On Mon, 2013-02-04 at 09:58 -0800, Thomas Pedersen wrote: >> On Mon, Feb 4, 2013 at 9:38 AM, Johannes Berg <johannes@sipsolutions.net> wrote: >> > On Sat, 2013-02-02 at 17:02 -0800, Thomas Pedersen wrote: >> > >> >> - pos = skb_put(skb, ie_ssid_len); >> >> - *pos++ = WLAN_EID_SSID; >> >> + if (ieee80211_vif_is_mesh(vif)) { >> >> + pos = skb_put(skb, 2 + 2 + ssid_len); >> >> + *pos++ = WLAN_EID_SSID; >> >> + *pos++ = 0; >> >> + /* NOTE: mesh ID will be out of order */ >> > >> > Why put it out of order? >> > >> > Also I'm not convinced that it's a good idea to translate "SSID" from >> > the userspace API to "mesh ID" silently? Might make more sense to have >> > those separately maybe? I mean, it seems reasonable to even think you >> > might scan for a mesh network when you're not a mesh interface, for >> > example? >> >> Yeah that makes more sense, but will obviously require more work. >> I'll just drop this for now. > > I guess the other question is if you actually want this at all. I mean, > if you just do patch 1 and 3, then unless you want to actively scan for > multiple networks you can just include the mesh ID in the IE parameter. > Seems for many purposes that could be acceptable already. Now I'm not > saying that you shouldn't put it into the kernel, there certainly could > be value in that, just saying that in terms of effort that might be > easier? Yeah, allowing userspace to specify the mesh ID IEs would be easy for now. Thanks,
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 01dc001..9f3e9ba 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -2622,11 +2622,18 @@ struct sk_buff *ieee80211_probereq_get(struct ieee80211_hw *hw, memcpy(hdr->addr2, vif->addr, ETH_ALEN); eth_broadcast_addr(hdr->addr3); - pos = skb_put(skb, ie_ssid_len); - *pos++ = WLAN_EID_SSID; + if (ieee80211_vif_is_mesh(vif)) { + pos = skb_put(skb, 2 + 2 + ssid_len); + *pos++ = WLAN_EID_SSID; + *pos++ = 0; + /* NOTE: mesh ID will be out of order */ + *pos++ = WLAN_EID_MESH_ID; + } else { + pos = skb_put(skb, 2 + ssid_len); + *pos++ = WLAN_EID_SSID; + } *pos++ = ssid_len; - if (ssid_len) - memcpy(pos, ssid, ssid_len); + memcpy(pos, ssid, ssid_len); pos += ssid_len; return skb; diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 139ad9b..aeb0f88 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -1181,6 +1181,7 @@ int ieee80211_build_preq_ies(struct ieee80211_local *local, u8 *buffer, if (ie && ie_len) { static const u8 before_ht[] = { WLAN_EID_SSID, + WLAN_EID_MESH_ID, WLAN_EID_SUPP_RATES, WLAN_EID_REQUEST, WLAN_EID_EXT_SUPP_RATES,
A mesh probe request must include the mesh ID IE and a wildcard SSID IE. Add this and transpose the SSID scan list into mesh IDs for mesh interfaces. Also allow the user to specify a mesh ID IE in the user scan IEs, since additional SSID IEs seem to be allowed. Signed-off-by: Thomas Pedersen <thomas@cozybit.com> --- net/mac80211/tx.c | 15 +++++++++++---- net/mac80211/util.c | 1 + 2 files changed, 12 insertions(+), 4 deletions(-)