)]}' { "log": [ { "commit": "9fbc630c89fd210e15ffe84fd6e968a2d39000b0", "tree": "7cd2283f43fcae0ce5779cffd976579977757989", "parents": [ "b9f2e39d4c2bcd8e94f73ae14450d7764f930a41" ], "author": { "name": "Felix Fietkau", "email": "nbd@openwrt.org", "time": "Sat May 15 15:46:17 2010 +0200" }, "committer": { "name": "John W. Linville", "email": "linville@tuxdriver.com", "time": "Fri May 21 14:40:00 2010 -0400" }, "message": "cfg80211: fix crash in cfg80211_set_freq()\n\nSince wdev can be NULL, check it before dereferencing it\n\nSigned-off-by: Felix Fietkau \u003cnbd@openwrt.org\u003e\nSigned-off-by: John W. Linville \u003clinville@tuxdriver.com\u003e\n" }, { "commit": "f444de05d20e27cdd960c13fcbcfca3099f03143", "tree": "a7fbef60420d88dda5840e06094be21ee3eb1dc0", "parents": [ "ac8dd506e40ee2c7fcc61654a44c32555a0a8d6c" ], "author": { "name": "Johannes Berg", "email": "johannes@sipsolutions.net", "time": "Wed May 05 15:25:02 2010 +0200" }, "committer": { "name": "John W. Linville", "email": "linville@tuxdriver.com", "time": "Fri May 07 14:55:50 2010 -0400" }, "message": "cfg80211/mac80211: better channel handling\n\nCurrently (all tested with hwsim) you can do stupid\nthings like setting up an AP on a certain channel,\nthen adding another virtual interface and making\nthat associate on another channel -- this will make\nthe beaconing to move channel but obviously without\nthe necessary IEs data update.\n\nIn order to improve this situation, first make the\nconfiguration APIs (cfg80211 and nl80211) aware of\nmulti-channel operation -- we\u0027ll eventually need\nthat in the future anyway. There\u0027s one userland API\nchange and one API addition. The API change is that\nnow SET_WIPHY must be called with virtual interface\nindex rather than only wiphy index in order to take\neffect for that interface -- luckily all current\nusers (hostapd) do that. For monitor interfaces, the\nold setting is preserved, but monitors are always\nslaved to other devices anyway so no guarantees.\n\nThe second userland API change is the introduction\nof a per virtual interface SET_CHANNEL command, that\nhostapd should use going forward to make it easier\nto understand what\u0027s going on (it can automatically\ndetect a kernel with this command).\n\nOther than mac80211, no existing cfg80211 drivers\nare affected by this change because they only allow\na single virtual interface.\n\nmac80211, however, now needs to be aware that the\nchannel settings are per interface now, and needs\nto disallow (for now) real multi-channel operation,\nwhich is another important part of this patch.\n\nOne of the immediate benefits is that you can now\nstart hostapd to operate on a hardware that already\nhas a connection on another virtual interface, as\nlong as you specify the same channel.\n\nNote that two things are left unhandled (this is an\nimprovement -- not a complete fix):\n\n * different HT/no-HT modes\n\n currently you could start an HT AP and then\n connect to a non-HT network on the same channel\n which would configure the hardware for no HT;\n that can be fixed fairly easily\n\n * CSA\n\n An AP we\u0027re connected to on a virtual interface\n might indicate switching channels, and in that\n case we would follow it, regardless of how many\n other interfaces are operating; this requires\n more effort to fix but is pretty rare after all\n\nSigned-off-by: Johannes Berg \u003cjohannes@sipsolutions.net\u003e\nSigned-off-by: John W. Linville \u003clinville@tuxdriver.com\u003e\n" }, { "commit": "9588bbd5529461a3dacd435bf239c84c3508f569", "tree": "dd00011714bdefdef1f818ac7caf53682253a515", "parents": [ "b203ffc3a447eb8d9e6120b783ddee081b143061" ], "author": { "name": "Jouni Malinen", "email": "jouni.malinen@atheros.com", "time": "Wed Dec 23 13:15:41 2009 +0100" }, "committer": { "name": "John W. Linville", "email": "linville@tuxdriver.com", "time": "Mon Dec 28 16:55:02 2009 -0500" }, "message": "cfg80211: add remain-on-channel command\n\nAdd new commands for requesting the driver to remain awake\non a specified channel for the specified amount of time\n(and another command to cancel such an operation). This\ncan be used to implement userspace-controlled off-channel\noperations, like Public Action frame exchange on another\nchannel than the operation channel.\n\nThe off-channel operation should behave similarly to scan,\ni.e. the local station (if associated) moves into power\nsave mode to request the AP to buffer frames for it and\nthen moves to the other channel to allow the off-channel\noperation to be completed. The duration parameter can be\nused to request enough time to receive a response from\nthe target station.\n\nSigned-off-by: Jouni Malinen \u003cjouni.malinen@atheros.com\u003e\nSigned-off-by: Johannes Berg \u003cjohannes@sipsolutions.net\u003e\nSigned-off-by: John W. Linville \u003clinville@tuxdriver.com\u003e\n" }, { "commit": "4b181144e6c1c25aaba9b9fc7cc70c95495ecb92", "tree": "93f85f1c99f097ec3c60f45b7b917c1b0dfbf7f9", "parents": [ "e312c24cf8229f9b6e76dbfd5d99eefe21f4ac0a" ], "author": { "name": "Johannes Berg", "email": "johannes@sipsolutions.net", "time": "Sat Aug 08 11:03:58 2009 +0200" }, "committer": { "name": "John W. Linville", "email": "linville@tuxdriver.com", "time": "Fri Aug 14 09:13:51 2009 -0400" }, "message": "cfg80211: fix locking for SIWFREQ\n\n\"cfg80211: validate channel settings across interfaces\"\ncontained a locking bug -- in the managed-mode SIWFREQ\ncall it would end up running into a lock recursion.\n\nThis fixes it by not checking that particular interface\nfor a channel that it needs to stay on, which is as it\nshould be as that\u0027s the interface we\u0027re setting the\nchannel for.\n\nReported-by: Reinette Chatre \u003creinette.chatre@intel.com\u003e\nReported-by: Kalle Valo \u003ckalle.valo@iki.fi\u003e\nSigned-off-by: Johannes Berg \u003cjohannes@sipsolutions.net\u003e\nTested-by: Kalle Valo \u003ckalle.valo@iki.fi\u003e\nTested-by: Reinette Chatre \u003creinette.chatre@intel.com\u003e\nSigned-off-by: John W. Linville \u003clinville@tuxdriver.com\u003e\n" }, { "commit": "59bbb6f7574bc693ed8313b98eac641116c95b94", "tree": "da24ed15c5e375782e79b3dab7022d2100a7987a", "parents": [ "f26b32ed4bd5780855a79bb17fb1a431fa867dad" ], "author": { "name": "Johannes Berg", "email": "johannes@sipsolutions.net", "time": "Fri Aug 07 17:22:35 2009 +0200" }, "committer": { "name": "John W. Linville", "email": "linville@tuxdriver.com", "time": "Fri Aug 14 09:13:42 2009 -0400" }, "message": "cfg80211: validate channel settings across interfaces\n\nCurrently, there\u0027s a problem that affects regulatory\nenforcement and connection stability, in that it is\npossible to switch the channel while connected to a\nnetwork or joined to an IBSS.\n\nThe problem comes from the fact that we only validate\nthe channel against the current interface\u0027s type, not\nagainst any other interface. Thus, you have any type\nof interface up, additionally bring up a monitor mode\ninterface and switch the channel on the monitor. This\nwill obviously also switch the channel on the other\ninterface.\n\nThe problem now is that if you do that while sending\nbeacons for IBSS mode, you can switch to a disabled\nchannel or a channel that doesn\u0027t allow beaconing.\nCombined with a managed mode interface connected to\nan AP instead of an IBSS interface, you can easily\nbreak the connection that way.\n\nTo fix this, this patch validates any channel change\nwith all available interfaces, and disallows such\nchanges on secondary interfaces if another interface\nis connected to an AP or joined to an IBSS.\n\nSigned-off-by: Johannes Berg \u003cjohannes@sipsolutions.net\u003e\nSigned-off-by: John W. Linville \u003clinville@tuxdriver.com\u003e\n" } ] }