NetBSD/sys/net80211/ieee80211_crypto.c

667 lines
19 KiB
C
Raw Normal View History

/* $NetBSD: ieee80211_crypto.c,v 1.23 2018/05/08 07:02:07 maxv Exp $ */
/*
* Copyright (c) 2001 Atsushi Onoe
* Copyright (c) 2002-2005 Sam Leffler, Errno Consulting
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* Alternatively, this software may be distributed under the terms of the
* GNU General Public License ("GPL") version 2 as published by the Free
* Software Foundation.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <sys/cdefs.h>
2003-09-14 05:14:54 +04:00
#ifdef __FreeBSD__
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
__FBSDID("$FreeBSD: src/sys/net80211/ieee80211_crypto.c,v 1.12 2005/08/08 18:46:35 sam Exp $");
#endif
#ifdef __NetBSD__
__KERNEL_RCSID(0, "$NetBSD: ieee80211_crypto.c,v 1.23 2018/05/08 07:02:07 maxv Exp $");
2003-09-14 05:14:54 +04:00
#endif
2015-08-25 01:21:26 +03:00
#ifdef _KERNEL_OPT
#include "opt_inet.h"
2015-08-25 01:21:26 +03:00
#endif
/*
* IEEE 802.11 generic crypto support.
*/
#include <sys/param.h>
#include <sys/mbuf.h>
#include <sys/socket.h>
#include <sys/sockio.h>
#include <sys/endian.h>
#include <sys/errno.h>
#include <sys/proc.h>
#include <sys/sysctl.h>
#include <net/if.h>
#include <net/if_media.h>
#include <net/if_arp.h>
#include <net/if_ether.h>
#include <net/if_llc.h>
#include <net80211/ieee80211_netbsd.h>
#include <net80211/ieee80211_var.h>
/*
* Table of registered cipher modules.
*/
static const struct ieee80211_cipher *ciphers[IEEE80211_CIPHER_MAX];
#ifdef INET
#include <netinet/in.h>
#include <net/if_ether.h>
#endif
static int _ieee80211_crypto_delkey(struct ieee80211com *,
struct ieee80211_key *);
/*
* Default "null" key management routines.
*/
static int
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
null_key_alloc(struct ieee80211com *ic, const struct ieee80211_key *k,
ieee80211_keyix *keyix, ieee80211_keyix *rxkeyix)
{
2005-07-27 02:52:47 +04:00
if (!(&ic->ic_nw_keys[0] <= k &&
k < &ic->ic_nw_keys[IEEE80211_WEP_NKID])) {
/*
* Not in the global key table, the driver should handle this
* by allocating a slot in the h/w key table/cache. In
* lieu of that return key slot 0 for any unicast key
* request. We disallow the request if this is a group key.
* This default policy does the right thing for legacy hardware
* with a 4 key table. It also handles devices that pass
* packets through untouched when marked with the WEP bit
* and key index 0.
*/
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
if (k->wk_flags & IEEE80211_KEY_GROUP)
return 0;
*keyix = 0; /* NB: use key index 0 for ucast key */
} else {
*keyix = k - ic->ic_nw_keys;
2005-07-27 02:52:47 +04:00
}
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
*rxkeyix = IEEE80211_KEYIX_NONE; /* XXX maybe *keyix? */
return 1;
}
static int
null_key_delete(struct ieee80211com *ic, const struct ieee80211_key *k)
{
return 1;
}
static int
null_key_set(struct ieee80211com *ic, const struct ieee80211_key *k,
const u_int8_t mac[IEEE80211_ADDR_LEN])
{
return 1;
}
static void
null_key_update(struct ieee80211com *ic)
{
;
}
/*
* Write-arounds for common operations.
*/
static __inline void
cipher_detach(struct ieee80211_key *key)
{
key->wk_cipher->ic_detach(key);
}
/*
* Wrappers for driver key management methods.
*/
static __inline int
dev_key_alloc(struct ieee80211com *ic, const struct ieee80211_key *key,
ieee80211_keyix *keyix, ieee80211_keyix *rxkeyix)
{
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
return ic->ic_crypto.cs_key_alloc(ic, key, keyix, rxkeyix);
}
static __inline int
dev_key_delete(struct ieee80211com *ic, const struct ieee80211_key *key)
{
return ic->ic_crypto.cs_key_delete(ic, key);
}
static __inline int
dev_key_set(struct ieee80211com *ic, const struct ieee80211_key *key,
const u_int8_t mac[IEEE80211_ADDR_LEN])
{
return ic->ic_crypto.cs_key_set(ic, key, mac);
}
/*
* Setup crypto support.
*/
void
ieee80211_crypto_attach(struct ieee80211com *ic)
{
struct ieee80211_crypto_state *cs = &ic->ic_crypto;
int i;
/* NB: we assume everything is pre-zero'd */
cs->cs_def_txkey = IEEE80211_KEYIX_NONE;
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
cs->cs_max_keyix = IEEE80211_WEP_NKID;
ciphers[IEEE80211_CIPHER_NONE] = &ieee80211_cipher_none;
for (i = 0; i < IEEE80211_WEP_NKID; i++)
ieee80211_crypto_resetkey(ic, &cs->cs_nw_keys[i],
IEEE80211_KEYIX_NONE);
/*
* Initialize the driver key support routines to noop entries.
* This is useful especially for the cipher test modules.
*/
cs->cs_key_alloc = null_key_alloc;
cs->cs_key_set = null_key_set;
cs->cs_key_delete = null_key_delete;
cs->cs_key_update_begin = null_key_update;
cs->cs_key_update_end = null_key_update;
}
/*
* Teardown crypto support.
*/
void
ieee80211_crypto_detach(struct ieee80211com *ic)
{
ieee80211_crypto_delglobalkeys(ic);
}
/*
* Register a crypto cipher module.
*/
void
ieee80211_crypto_register(const struct ieee80211_cipher *cip)
{
if (cip->ic_cipher >= IEEE80211_CIPHER_MAX) {
printf("%s: cipher %s has an invalid cipher index %u\n",
__func__, cip->ic_name, cip->ic_cipher);
return;
}
if (ciphers[cip->ic_cipher] != NULL && ciphers[cip->ic_cipher] != cip) {
printf("%s: cipher %s registered with a different template\n",
__func__, cip->ic_name);
return;
}
ciphers[cip->ic_cipher] = cip;
}
/*
* Unregister a crypto cipher module.
*/
void
ieee80211_crypto_unregister(const struct ieee80211_cipher *cip)
{
if (cip->ic_cipher >= IEEE80211_CIPHER_MAX) {
printf("%s: cipher %s has an invalid cipher index %u\n",
__func__, cip->ic_name, cip->ic_cipher);
return;
}
if (ciphers[cip->ic_cipher] != NULL && ciphers[cip->ic_cipher] != cip) {
printf("%s: cipher %s registered with a different template\n",
__func__, cip->ic_name);
return;
}
/* NB: don't complain about not being registered */
/* XXX disallow if references */
ciphers[cip->ic_cipher] = NULL;
}
int
ieee80211_crypto_available(u_int cipher)
{
return cipher < IEEE80211_CIPHER_MAX && ciphers[cipher] != NULL;
}
/* XXX well-known names! */
static const char *cipher_modnames[] = {
"wlan_wep", /* IEEE80211_CIPHER_WEP */
"wlan_tkip", /* IEEE80211_CIPHER_TKIP */
"wlan_aes_ocb", /* IEEE80211_CIPHER_AES_OCB */
"wlan_ccmp", /* IEEE80211_CIPHER_AES_CCM */
"wlan_ckip", /* IEEE80211_CIPHER_CKIP */
};
/*
* Establish a relationship between the specified key and cipher
* and, if necessary, allocate a hardware index from the driver.
* Note that when a fixed key index is required it must be specified
* and we blindly assign it w/o consulting the driver (XXX).
*
* This must be the first call applied to a key; all the other key
* routines assume wk_cipher is setup.
*
* Locking must be handled by the caller using:
* ieee80211_key_update_begin(ic);
* ieee80211_key_update_end(ic);
*/
int
ieee80211_crypto_newkey(struct ieee80211com *ic, int cipher, int flags,
struct ieee80211_key *key)
{
#define N(a) (sizeof(a) / sizeof(a[0]))
const struct ieee80211_cipher *cip;
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
ieee80211_keyix keyix, rxkeyix;
void *keyctx;
int oflags;
/*
* Validate cipher and set reference to cipher routines.
*/
if (cipher >= IEEE80211_CIPHER_MAX) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: invalid cipher %u\n", __func__, cipher);
ic->ic_stats.is_crypto_badcipher++;
return 0;
}
cip = ciphers[cipher];
if (cip == NULL) {
/*
* Auto-load cipher module if we have a well-known name
* for it. It might be better to use string names rather
* than numbers and craft a module name based on the cipher
* name; e.g. wlan_cipher_<cipher-name>.
*/
if (cipher < N(cipher_modnames)) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: unregistered cipher %u, load module %s\n",
__func__, cipher, cipher_modnames[cipher]);
ieee80211_load_module(cipher_modnames[cipher]);
/*
* If cipher module loaded it should immediately
* call ieee80211_crypto_register which will fill
* in the entry in the ciphers array.
*/
cip = ciphers[cipher];
}
if (cip == NULL) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: unable to load cipher %u, module %s\n",
__func__, cipher,
cipher < N(cipher_modnames) ?
cipher_modnames[cipher] : "<unknown>");
ic->ic_stats.is_crypto_nocipher++;
return 0;
}
}
oflags = key->wk_flags;
flags &= IEEE80211_KEY_COMMON;
/*
* If the hardware does not support the cipher then
* fall back to a host-based implementation.
*/
if ((ic->ic_caps & (1<<cipher)) == 0) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: no h/w support for cipher %s, falling back to s/w\n",
__func__, cip->ic_name);
flags |= IEEE80211_KEY_SWCRYPT;
}
/*
* Hardware TKIP with software MIC is an important
* combination; we handle it by flagging each key,
* the cipher modules honor it.
*/
if (cipher == IEEE80211_CIPHER_TKIP &&
(ic->ic_caps & IEEE80211_C_TKIPMIC) == 0) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: no h/w support for TKIP MIC, falling back to s/w\n",
__func__);
flags |= IEEE80211_KEY_SWMIC;
}
/*
* Bind cipher to key instance. Note we do this
* after checking the device capabilities so the
* cipher module can optimize space usage based on
* whether or not it needs to do the cipher work.
*/
if (key->wk_cipher != cip || key->wk_flags != flags) {
again:
/*
* Fillin the flags so cipher modules can see s/w
* crypto requirements and potentially allocate
* different state and/or attach different method
* pointers.
*
* XXX this is not right when s/w crypto fallback
* fails and we try to restore previous state.
*/
key->wk_flags = flags;
keyctx = cip->ic_attach(ic, key);
if (keyctx == NULL) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: unable to attach cipher %s\n",
__func__, cip->ic_name);
key->wk_flags = oflags; /* restore old flags */
ic->ic_stats.is_crypto_attachfail++;
return 0;
}
cipher_detach(key);
key->wk_cipher = cip; /* XXX refcnt? */
key->wk_private = keyctx;
}
/*
* Commit to requested usage so driver can see the flags.
*/
key->wk_flags = flags;
/*
* Ask the driver for a key index if we don't have one.
* Note that entries in the global key table always have
* an index; this means it's safe to call this routine
* for these entries just to setup the reference to the
* cipher template. Note also that when using software
* crypto we also call the driver to give us a key index.
*/
if (key->wk_keyix == IEEE80211_KEYIX_NONE) {
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
if (!dev_key_alloc(ic, key, &keyix, &rxkeyix)) {
/*
* Driver has no room; fallback to doing crypto
* in the host. We change the flags and start the
* procedure over. If we get back here then there's
* no hope and we bail. Note that this can leave
* the key in a inconsistent state if the caller
* continues to use it.
*/
if ((key->wk_flags & IEEE80211_KEY_SWCRYPT) == 0) {
ic->ic_stats.is_crypto_swfallback++;
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: no h/w resources for cipher %s, "
"falling back to s/w\n", __func__,
cip->ic_name);
oflags = key->wk_flags;
flags |= IEEE80211_KEY_SWCRYPT;
if (cipher == IEEE80211_CIPHER_TKIP)
flags |= IEEE80211_KEY_SWMIC;
goto again;
}
ic->ic_stats.is_crypto_keyfail++;
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: unable to setup cipher %s\n",
__func__, cip->ic_name);
return 0;
}
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
key->wk_keyix = keyix;
key->wk_rxkeyix = rxkeyix;
}
return 1;
#undef N
}
/*
* Remove the key (no locking, for internal use).
*/
static int
_ieee80211_crypto_delkey(struct ieee80211com *ic, struct ieee80211_key *key)
{
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
ieee80211_keyix keyix;
IASSERT(key->wk_cipher != NULL, ("No cipher!"));
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: %s keyix %u flags 0x%x rsc %ju tsc %ju len %u\n",
__func__, key->wk_cipher->ic_name,
key->wk_keyix, key->wk_flags,
key->wk_keyrsc, key->wk_keytsc, key->wk_keylen);
keyix = key->wk_keyix;
if (keyix != IEEE80211_KEYIX_NONE) {
/*
* Remove hardware entry.
*/
/* XXX key cache */
if (!dev_key_delete(ic, key)) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: driver did not delete key index %u\n",
__func__, keyix);
ic->ic_stats.is_crypto_delkey++;
/* XXX recovery? */
}
}
cipher_detach(key);
memset(key, 0, sizeof(*key));
ieee80211_crypto_resetkey(ic, key, IEEE80211_KEYIX_NONE);
return 1;
}
/*
* Remove the specified key.
*/
int
ieee80211_crypto_delkey(struct ieee80211com *ic, struct ieee80211_key *key)
{
int status;
ieee80211_key_update_begin(ic);
status = _ieee80211_crypto_delkey(ic, key);
ieee80211_key_update_end(ic);
return status;
}
/*
* Clear the global key table.
*/
void
ieee80211_crypto_delglobalkeys(struct ieee80211com *ic)
{
int i;
ieee80211_key_update_begin(ic);
for (i = 0; i < IEEE80211_WEP_NKID; i++)
(void)_ieee80211_crypto_delkey(ic, &ic->ic_nw_keys[i]);
ieee80211_key_update_end(ic);
}
/*
* Set the contents of the specified key.
*
* Locking must be handled by the caller using:
* ieee80211_key_update_begin(ic);
* ieee80211_key_update_end(ic);
*/
int
ieee80211_crypto_setkey(struct ieee80211com *ic, struct ieee80211_key *key,
const u_int8_t macaddr[IEEE80211_ADDR_LEN])
{
const struct ieee80211_cipher *cip = key->wk_cipher;
IASSERT(cip != NULL, ("No cipher!"));
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: %s keyix %u flags 0x%x mac %s rsc %ju tsc %ju len %u\n",
__func__, cip->ic_name, key->wk_keyix,
key->wk_flags, ether_sprintf(macaddr),
key->wk_keyrsc, key->wk_keytsc, key->wk_keylen);
/*
* Give cipher a chance to validate key contents.
* XXX should happen before modifying state.
*/
if (!cip->ic_setkey(key)) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: cipher %s rejected key index %u len %u flags 0x%x\n",
__func__, cip->ic_name, key->wk_keyix,
key->wk_keylen, key->wk_flags);
ic->ic_stats.is_crypto_setkey_cipher++;
return 0;
}
if (key->wk_keyix == IEEE80211_KEYIX_NONE) {
/* XXX nothing allocated, should not happen */
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"%s: no key index; should not happen!\n", __func__);
ic->ic_stats.is_crypto_setkey_nokey++;
return 0;
}
return dev_key_set(ic, key, macaddr);
}
/*
* Add privacy headers appropriate for the specified key.
*
* XXX XXX XXX: Here we modify 'm', and potentially reallocate it. We
* should pass back to the caller the updated pointer to avoid
* use-after-frees. This can be done by changing the argument to be **m,
* but many drivers will have to be changed accordingly.
*/
struct ieee80211_key *
ieee80211_crypto_encap(struct ieee80211com *ic, struct ieee80211_node *ni,
struct mbuf *m)
{
struct ieee80211_key *k;
struct ieee80211_frame *wh;
const struct ieee80211_cipher *cip;
u_int8_t keyid, *hdr;
int hdrlen;
/*
* Multicast traffic always uses the multicast key.
* Otherwise if a unicast key is set we use that and
* it is always key index 0. When no unicast key is
* set we fall back to the default transmit key.
*/
wh = mtod(m, struct ieee80211_frame *);
if (IEEE80211_IS_MULTICAST(wh->i_addr1) ||
ni->ni_ucastkey.wk_cipher == &ieee80211_cipher_none) {
if (ic->ic_def_txkey == IEEE80211_KEYIX_NONE) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_CRYPTO,
"[%s] no default transmit key (%s) deftxkey %u\n",
ether_sprintf(wh->i_addr1), __func__,
ic->ic_def_txkey);
ic->ic_stats.is_tx_nodefkey++;
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
return NULL;
}
keyid = ic->ic_def_txkey;
k = &ic->ic_nw_keys[ic->ic_def_txkey];
} else {
keyid = 0;
k = &ni->ni_ucastkey;
}
cip = k->wk_cipher;
/*
* The crypto header is added after the IEEE802.11 header. Prepend
* the size of the crypto header, and move the IEEE802.11 header back
* to the beginning of the mbuf. Ensure everything is contiguous.
*/
hdrlen = ieee80211_hdrspace(ic, mtod(m, void *));
M_PREPEND(m, cip->ic_header, M_NOWAIT);
if (m && m->m_len < hdrlen + cip->ic_header) {
m = m_pullup(m, hdrlen + cip->ic_header);
}
if (m == NULL) {
return NULL;
}
hdr = mtod(m, u_int8_t *);
memmove(hdr, hdr + cip->ic_header, hdrlen);
Resolve conflicts and adapt to NetBSD. Thanks to dyoung@, scw@, and perry@ for help testing. 2005-08-30 15:27 avatar Properly set ic_curchan before calling back to device driver to do channel switching(ifconfig devX channel Y). This fix should make channel changing works again in monitor mode. Submitted by: sam X-MFC-With: other ic_curchan changes 2005-08-13 18:50 sam revert 1.64: we cannot use the channel characteristics to decide when to do 11g erp sta accounting because b/g channels show up as false positives when operating in 11b. Noticed by: Michal Mertl 2005-08-13 18:31 sam Extend acl support to pass ioctl requests through and use this to add support for getting the current policy setting and collecting the list of mac addresses in the acl table. Submitted by: Michal Mertl (original version) MFC after: 2 weeks 2005-08-10 18:42 sam Don't use ic_curmode to decide when to do 11g station accounting, use the station channel properties. Fixes assert failure/bogus operation when an ap is operating in 11a and has associated stations then switches to 11g. Noticed by: Michal Mertl Reviewed by: avatar MFC after: 2 weeks 2005-08-10 17:22 sam Clarify/fix handling of the current channel: o add ic_curchan and use it uniformly for specifying the current channel instead of overloading ic->ic_bss->ni_chan (or in some drivers ic_ibss_chan) o add ieee80211_scanparams structure to encapsulate scanning-related state captured for rx frames o move rx beacon+probe response frame handling into separate routines o change beacon+probe response handling to treat the scan table more like a scan cache--look for an existing entry before adding a new one; this combined with ic_curchan use corrects handling of stations that were previously found at a different channel o move adhoc neighbor discovery by beacon+probe response frames to a new ieee80211_add_neighbor routine Reviewed by: avatar Tested by: avatar, Michal Mertl MFC after: 2 weeks 2005-08-09 11:19 rwatson Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to ifnet.if_drv_flags. Device drivers are now responsible for synchronizing access to these flags, as they are in if_drv_flags. This helps prevent races between the network stack and device driver in maintaining the interface flags field. Many __FreeBSD__ and __FreeBSD_version checks maintained and continued; some less so. Reviewed by: pjd, bz MFC after: 7 days 2005-08-08 19:46 sam Split crypto tx+rx key indices and add a key index -> node mapping table: Crypto changes: o change driver/net80211 key_alloc api to return tx+rx key indices; a driver can leave the rx key index set to IEEE80211_KEYIX_NONE or set it to be the same as the tx key index (the former disables use of the key index in building the keyix->node mapping table and is the default setup for naive drivers by null_key_alloc) o add cs_max_keyid to crypto state to specify the max h/w key index a driver will return; this is used to allocate the key index mapping table and to bounds check table loookups o while here introduce ieee80211_keyix (finally) for the type of a h/w key index o change crypto notifiers for rx failures to pass the rx key index up as appropriate (michael failure, replay, etc.) Node table changes: o optionally allocate a h/w key index to node mapping table for the station table using the max key index setting supplied by drivers (note the scan table does not get a map) o defer node table allocation to lateattach so the driver has a chance to set the max key id to size the key index map o while here also defer the aid bitmap allocation o add new ieee80211_find_rxnode_withkey api to find a sta/node entry on frame receive with an optional h/w key index to use in checking mapping table; also updates the map if it does a hash lookup and the found node has a rx key index set in the unicast key; note this work is separated from the old ieee80211_find_rxnode call so drivers do not need to be aware of the new mechanism o move some node table manipulation under the node table lock to close a race on node delete o add ieee80211_node_delucastkey to do the dirty work of deleting unicast key state for a node (deletes any key and handles key map references) Ath driver: o nuke private sc_keyixmap mechansim in favor of net80211 support o update key alloc api These changes close several race conditions for the ath driver operating in ap mode. Other drivers should see no change. Station mode operation for ath no longer uses the key index map but performance tests show no noticeable change and this will be fixed when the scan table is eliminated with the new scanning support. Tested by: Michal Mertl, avatar, others Reviewed by: avatar, others MFC after: 2 weeks 2005-08-08 06:49 sam use ieee80211_iterate_nodes to retrieve station data; the previous code walked the list w/o locking MFC after: 1 week 2005-08-08 04:30 sam Cleanup beacon/listen interval handling: o separate configured beacon interval from listen interval; this avoids potential use of one value for the other (e.g. setting powersavesleep to 0 clobbers the beacon interval used in hostap or ibss mode) o bounds check the beacon interval received in probe response and beacon frames and drop frames with bogus settings; not clear if we should instead clamp the value as any alteration would result in mismatched sta+ap configuration and probably be more confusing (don't want to log to the console but perhaps ok with rate limiting) o while here up max beacon interval to reflect WiFi standard Noticed by: Martin <nakal@nurfuerspam.de> MFC after: 1 week 2005-08-06 05:57 sam fix debug msg typo MFC after: 3 days 2005-08-06 05:56 sam Fix handling of frames sent prior to a station being authorized when operating in ap mode. Previously we allocated a node from the station table, sent the frame (using the node), then released the reference that "held the frame in the table". But while the frame was in flight the node might be reclaimed which could lead to problems. The solution is to add an ieee80211_tmp_node routine that crafts a node that does exist in a table and so isn't ever reclaimed; it exists only so long as the associated frame is in flight. MFC after: 5 days 2005-07-31 07:12 sam close a race between reclaiming a node when a station is inactive and sending the null data frame used to probe inactive stations MFC after: 5 days 2005-07-27 05:41 sam when bridging internally bypass the bss node as traffic to it must follow the normal input path Submitted by: Michal Mertl MFC after: 5 days 2005-07-27 03:53 sam bandaid ni_fails handling so ap's with association failures are reconsidered after a bit; a proper fix involves more changes to the scanning infrastructure Reviewed by: avatar, David Young MFC after: 5 days 2005-07-23 01:16 sam the AREF flag is only meaningful in ap mode; adhoc neighbors now are timed out of the sta/neighbor table 2005-07-23 00:25 sam o move inactivity-related debug msgs under IEEE80211_MSG_INACT o probe inactive neighbors in adhoc mode (they don't have an association id so previously were being timed out) MFC after: 3 days 2005-07-22 22:11 sam split xmit of probe request frame out into a separate routine that takes explicit parameters; this will be needed when scanning is decoupled from the state machine to do bg scanning MFC after: 3 days 2005-07-22 21:48 sam split 802.11 frame xmit setup code into ieee80211_send_setup MFC after: 3 days 2005-07-22 18:57 sam simplify ic_newassoc callback MFC after: 3 days 2005-07-22 18:54 sam simplify ieee80211_ibss_merge api MFC after: 3 days 2005-07-22 18:50 sam add stats we know we'll need soon and some spare fields for future expansion MFC after: 3 days 2005-07-22 18:45 sam simplify tim callback api MFC after: 3 days 2005-07-22 18:42 sam don't include 802.3 header in min frame length calculation as it may not be present for a frag; fixes problem with small (fragmented) frames being dropped Obtained from: Atheros MFC after: 3 days 2005-07-22 18:36 sam simplify ieee80211_node_authorize and ieee80211_node_unauthorize api's MFC after: 3 days 2005-07-22 18:31 sam simplifiy ieee80211_send_nulldata api MFC after: 3 days 2005-07-22 18:29 sam simplify rate set api's by removing ic parameter (implicit in node reference) MFC after: 3 days 2005-07-22 18:21 sam reject association requests with a wpa/rsn ie when wpa/rsn is not configured on the ap; previously we either ignored the ie or (possibly) failed an assertion Obtained from: Atheros MFC after: 3 days 2005-07-22 18:16 sam missed one in last commit; add device name to discard msgs 2005-07-22 18:13 sam include device name in discard msgs 2005-07-22 18:12 sam add diag msgs for frames discarded because the direction field is wrong 2005-07-22 18:08 sam split data frame delivery out to a new function ieee80211_deliver_data 2005-07-22 18:00 sam o add IEEE80211_IOC_FRAGTHRESHOLD for getting+setting the tx fragmentation threshold o fix bounds checking on IEEE80211_IOC_RTSTHRESHOLD MFC after: 3 days 2005-07-22 17:55 sam o add IEEE80211_FRAG_DEFAULT o move default settings for RTS and frag thresholds to ieee80211_var.h 2005-07-22 17:50 sam diff reduction against p4: define IEEE80211_FIXED_RATE_NONE and use it instead of -1 2005-07-22 17:37 sam add flags missed in last merge 2005-07-22 17:36 sam Diff reduction against p4: o add ic_flags_ext for eventual extention of ic_flags o define/reserve flag+capabilities bits for superg, bg scan, and roaming support o refactor debug msg macros MFC after: 3 days 2005-07-22 06:17 sam send a response when an auth request is denied due to an acl; might be better to silently ignore the frame but this way we give stations a chance of figuring out what's wrong 2005-07-22 06:15 sam remove excess whitespace 2005-07-22 05:55 sam use IF_HANDOFF when bridging frames internally so if_start gets called; fixes communication between associated sta's MFC after: 3 days 2005-07-11 04:06 sam Handle encrypt of arbitarily fragmented mbuf chains: previously we bailed if we couldn't collect the 16-bytes of data required for an aes block cipher in 2 mbufs; now we deal with it. While here make space accounting signed so a sanity check does the right thing for malformed mbuf chains. Approved by: re (scottl) 2005-07-11 04:00 sam nuke assert that duplicates real check Reviewed by: avatar Approved by: re (scottl)
2005-11-18 19:40:08 +03:00
return (cip->ic_encap(k, m, keyid<<6) ? k : NULL);
}
#define IEEE80211_WEP_HDRLEN (IEEE80211_WEP_IVLEN + IEEE80211_WEP_KIDLEN)
#define IEEE80211_WEP_MINLEN \
(sizeof(struct ieee80211_frame) + \
IEEE80211_WEP_HDRLEN + IEEE80211_WEP_CRCLEN)
/*
* Validate and strip privacy headers (and trailer) for a
* received frame that has the WEP/Privacy bit set.
*/
struct ieee80211_key *
ieee80211_crypto_decap(struct ieee80211com *ic,
struct ieee80211_node *ni, struct mbuf **mp, int hdrlen)
{
const struct ieee80211_cipher *cip;
struct ieee80211_key *k;
struct ieee80211_frame *wh;
struct mbuf *m = *mp;
u_int8_t keyid;
KASSERT((m->m_flags & M_PKTHDR) != 0);
/*
* This minimum size data frame could be bigger. It is re-checked
* below.
*/
if (m->m_pkthdr.len < IEEE80211_WEP_MINLEN) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_ANY,
"%s: WEP data frame too short, len %u\n",
__func__, m->m_pkthdr.len);
ic->ic_stats.is_rx_tooshort++;
return NULL;
}
/*
* Locate the key. If unicast and there is no unicast
* key then we fall back to the key id in the header.
* This assumes unicast keys are only configured when
* the key id in the header is meaningless (typically 0).
*/
wh = mtod(m, struct ieee80211_frame *);
m_copydata(m, hdrlen + IEEE80211_WEP_IVLEN, sizeof(keyid), &keyid);
if (IEEE80211_IS_MULTICAST(wh->i_addr1) ||
ni->ni_ucastkey.wk_cipher == &ieee80211_cipher_none) {
k = &ic->ic_nw_keys[keyid >> 6];
} else {
k = &ni->ni_ucastkey;
}
/*
* Insure crypto header is contiguous for all decap work.
*/
cip = k->wk_cipher;
if (m->m_len < hdrlen + cip->ic_header) {
m = m_pullup(m, hdrlen + cip->ic_header);
*mp = m;
}
if (m == NULL) {
ic->ic_stats.is_rx_tooshort++;
return NULL;
}
/*
* Ensure there is a header+trailer included.
*/
if (m->m_pkthdr.len < hdrlen + cip->ic_header + cip->ic_trailer) {
IEEE80211_DPRINTF(ic, IEEE80211_MSG_ANY,
"%s: WEP data frame too short, len %u\n",
__func__, m->m_pkthdr.len);
ic->ic_stats.is_rx_tooshort++;
return NULL;
}
2005-07-27 02:52:47 +04:00
return (cip->ic_decap(k, m, hdrlen) ? k : NULL);
}