- 25 Nov, 2013 40 commits
-
-
Luis R. Rodriguez authored
last_request is RCU protected, since we're getting it on set_regdom() we might as well pass it to ensure the same request is being processed, otherwise there is a small race it could have changed. This makes processing of the request atomic. Signed-off-by:
Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by:
Johannes Berg <johannes.berg@intel.com>
-
Luis R. Rodriguez authored
Drivers that set the WIPHY_FLAG_CUSTOM_REGULATORY skip the core world regulatory domain updates, but do want their reg_notifier() called. Move the check for this closer to the source of the check that detected skipped was required and while at it add a helper for the notifier calling. This has no functional changes. This brings together the place where we call the reg_notifier() will be called. Signed-off-by:
-
Luis R. Rodriguez authored
It seems some out of tree drivers were using a regulatory_hint("00") to trigger off the wiphy regulatory notifier, for those cases just setting the WIPHY_FLAG_CUSTOM_REGULATORY would suffice to call the reg_notifier() for a world regulatory domain update. If drivers find other needs for calling the reg_notifier() a proper implemenation is preferred. Signed-off-by: -
Luis R. Rodriguez authored
All the regulatory request process routines use the same pattern. Signed-off-by:
-
Luis R. Rodriguez authored
This enforces proper RCU APIs accross the code. Signed-off-by:
-
Luis R. Rodriguez authored
This is common code, this reduces the chance of making a mistake of how we free it. Signed-off-by:
-
Luis R. Rodriguez authored
By dealing with non country IE conficts first we can shift the code that deals with the conflict to the left. This has no functional changes. Signed-off-by:
-
Luis R. Rodriguez authored
This is the last split up of the old unified __regultory_hint() processing set of functionality, it moves the country IE processing all on its own. This makes it easier to follow and read what exactly is going on for the case of processing country IEs. Signed-off-by:
-
Luis R. Rodriguez authored
This makes the code easier to read and follow. Signed-off-by:
-
Luis R. Rodriguez authored
This makes the code path easier to read and lets us split out some functionality that is only user specific, that makes it easier to read the other types of requests. Signed-off-by:
-
Luis R. Rodriguez authored
This makes the code path easier to read for the core case. Signed-off-by:
-
Luis R. Rodriguez authored
Currently nl80211 allows userspace to send the kernel a bogus regulatory domain with at most 32 rules set and it won't reject it until after its allocated memory. Let's be smart about it and take advantage that the last_request is now available under RTNL and check if the alpha2 matches an expected request and reject any bogus userspace requests prior to hitting the memory allocator. Signed-off-by:
-
Luis R. Rodriguez authored
If a custom regulatory domain is passed and if a rule for a channel indicates it should be disabled that channel should always remain disabled as per its documentation and design. Likewise if WIPHY_FLAG_STRICT_REGULATORY flag is set and a regulatory_hint() is issued if a channel is disabled that channel should remain disabled. Without this change only drivers that set the _orig flags appropriately on their own would ensure disallowed channels remaind disabled. This helps drivers save code by relying on the APIS provided to entrust channels that should not be enabled be respected by only having to use wiphy_apply_custom_regulatory() or regulatory_hint() with the WIPHY_FLAG_STRICT_REGULATORY set. If wiphy_apply_custom_regulatory() is used together with WIPHY_FLAG_STRICT_REGULATORY and a regulatory_hint() issued later, the incoming regulatory domain can override previously set _orig parameters from the initial custom regulatory setting. Signed-off-by:
-
Luis R. Rodriguez authored
If genregdb.awk assumes the file will end with an extra empty line or a comment line. This is could not be true so just address this. Signed-off-by:
John W. Linville <linville@tuxdriver.com> Signed-off-by:
-
Luis R. Rodriguez authored
This has no functional change, this just lets us reuse helpers at a later time. Signed-off-by:
-
Max Stepanov authored
This adds generic cipher scheme support to mac80211, such schemes are fully under control by the driver. On hw registration drivers may specify additional HW ciphers with a scheme how these ciphers have to be handled by mac80211 TX/RR. A cipher scheme specifies a cipher suite value, a size of the security header to be added to or stripped from frames and how the PN is to be verified on RX. Signed-off-by:
Max Stepanov <Max.Stepanov@intel.com> Signed-off-by:
-
Janusz Dziedzic authored
Allow beconing after we pass Channel Availability Check (CAC). Allow non-DFS and DFS channels mix. All DFS channels have to be in NL80211_DFS_AVAILABLE state (pass CAC). Signed-off-by:
Janusz Dziedzic <janusz.dziedzic@tieto.com> Signed-off-by:
-
Janusz Dziedzic authored
To report channel width correctly we have to send correct channel parameters from mac80211 when calling cfg80211_cac_event(). This is required in case of using channel width higher than 20MHz and we have to set correct dfs channel state after CAC (NL80211_DFS_AVAILABLE). Signed-off-by:
-
Johannes Berg authored
There's no code calling ieee80211_key_replace() with both arguments NULL and it wouldn't make sense, but in the interest of maintainability add a warning for it. As a side effect, this also shuts up a smatch warning. Signed-off-by: -
Johannes Berg authored
As the flag is entirely implemented in cfg80211, it should have been a global feature flag (which I believe didn't exist at the time). However, there's no reason to allow drivers to unset the flag, so don't allow it and remove the validation of NL80211_SCAN_FLAG_FLUSH. Signed-off-by: -
Johannes Berg authored
Improve readability of the function by adding the break, there's no functional impact but it's confusing to fall through. Signed-off-by: -
Johannes Berg authored
Coverity points out that checking assoc_data->ie is completely useless since it's an array in the struct and can't be NULL - remove the useless checks. Signed-off-by: -
Thomas Pedersen authored
802.11-2012 13.3.1 implicitly limits the mesh local link ID range to that of AID, since for mesh PS the local link ID must be indicated in the TIM IE, which only holds IEEE80211_MAX_AID bits. Also the code was allowing a local link ID of 0, but this is not correct since that TIM bit is used for indicating buffered mcast frames. Generate a random, unique, link ID from 1 - 2007, and drop a modulo conversion for the local link ID, but keep it for the peer link ID in case he chose something > MAX_AID. Signed-off-by:
Thomas Pedersen <thomas@cozybit.com> Signed-off-by:
-
Thomas Pedersen authored
Signed-off-by:
-
Thomas Pedersen authored
Signed-off-by:
-
Thomas Pedersen authored
If we store the peer link ID right after initializing a new neighbor, there is no need to do it later in the peering FSM. Signed-off-by:
-
Thomas Pedersen authored
Signed-off-by:
-
Thomas Pedersen authored
Signed-off-by:
-
Bob Copeland authored
Said function was removed some time ago. Signed-off-by:
Bob Copeland <bob@cozybit.com> Signed-off-by:
-
Bob Copeland authored
All other paths return an error code, do the same here. Signed-off-by:
-
Bob Copeland authored
The ignore_plink_timer flag is set when doing mod_timer() if the timer was not previously active. This is to avoid executing the timeout if del_timer() was subsequently called. However, del_timer() only happens if we are moving to ESTAB state or get a close frame while in HOLDING. We cannot leave HOLDING and re-enter ESTAB unless we receive a close frame (in which case ignore_plink_timer is already set) or if the timeout expires, so there actually isn't a case where this is needed on mod_timer(). Signed-off-by:
-
Bob Copeland authored
The matches_local check can just be done when looking at the individual action types. Signed-off-by:
-
Bob Copeland authored
Use C instead of cpp for type checking. Also swap the arguments into the usual sdata -> sta order. Signed-off-by:
-
Bob Copeland authored
The initial frame checks differ depending on whether this is a new peer or not, but they were all intermixed with sta checks as necessary. Group them together so the two cases are clearer. Signed-off-by:
-
Bob Copeland authored
Reject and accepted close events always put the host in the holding state and compute a reason code based only on the current state. Likewise on establish we always do the same setup. Put these in functions to save some duplicated code. Signed-off-by:
-
Bob Copeland authored
Rather than unlock at the end of each case, do it once after all is said and done. Signed-off-by:
-
Thomas Pedersen authored
Do all frame transfers in one place at the end of the big switch statements. sta->plid and sta->reason can be passed in any case, since they are only used for the frames that need them. Remove assignments to locals for values already stored in the sta structure. Signed-off-by:
Bob Copeland <me@bobcopeland.com> Signed-off-by:
-
Bob Copeland authored
According to IEEE 802.11-2012 (8.4.2.104), no peering management element exists with length 7. This code is checking to see if llid is present to ignore close frames with different llid, which would be IEs with length 8. Signed-off-by:
-
Luis R. Rodriguez authored
The iniator is already available to us, so use it. Signed-off-by:
-
Luis R. Rodriguez authored
wiphy_apply_custom_regulatory() implies WIPHY_FLAG_CUSTOM_REGULATORY but we never enforced it, do that now and warn if the driver didn't set it. All drivers should be following this today already. Having WIPHY_FLAG_CUSTOM_REGULATORY does not however mean you will use wiphy_apply_custom_regulatory() though, you may have your own _orig value set up tools / helpers. The intel drivers are examples of this type of driver. Signed-off-by:
-
