- 13 Feb, 2009 40 commits
-
-
Johannes Berg authored
Essentially consisting of passing the sta_info pointer around, instead of repeatedly doing hash lookups. Signed-off-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Johannes Berg authored
Clean up the locking by splitting it into two functions, this will also enable further cleanups of stopping all sessions. Signed-off-by:
-
Johannes Berg authored
The sta_info pointer can very well be passed to ieee80211_sta_tear_down_BA_sessions, this will later allow us to pass it through even further. Signed-off-by:
-
Johannes Berg authored
As far as I can tell, there are possible lockups because both the RX session_timer and TX addba_resp_timer are del_timer_sync'ed under the sta spinlock which both timer functions take. Additionally, the TX agg code seems to leak memory when TX aggregation is not disabled before the sta_info is freed. Fix this by making the free code a little smarter in the RX agg case, and actually make the sta_info_destroy code free the TX agg info in the TX agg case. We won't notify the peer, but it'll notice something is wrong anyway, and normally this only happens after we've told it in some other way we will no longer talk to it. Signed-off-by:
-
Johannes Berg authored
When disabling TX aggregation because it was rejected or from the timer (it was not accepted), there is a window where we first set the state to operation, unlock, and then undo the whole thing. Avoid that by splitting up the stop function. Also get rid of the pointless sta_info indirection in the timer. Signed-off-by:
-
Johannes Berg authored
Add documentation and move ieee80211_start_tx_ba_cb_irqsafe to right after ieee80211_start_tx_ba_cb. Signed-off-by:
-
Johannes Berg authored
Doing so would be an MLME protocol violation when the peer disabled the aggregation session. Quick driver review indicates that there are error codes passed all over the drivers but cannot ever be nonzero except in error conditions that would indicate mac80211 bugs. No real changes here, since no drivers currently can return -EBUSY. Signed-off-by:
-
Johannes Berg authored
We can only support aggregation on AP/STA right now. HT isn't defined for IBSS, WDS or MESH. In the WDS/MESH cases it's not clear what to put into the IBSS field, and we don't handle that in the code at all. Also fix the code to handle VLAN correctly. Signed-off-by:
-
Johannes Berg authored
Create two new files, agg-tx.c and agg-rx.c to make it clearer which code is common (ht.c) and which is specific (agg-*.c). Signed-off-by:
-
Johannes Berg authored
The values are in TUs (1.024ms), not ms. Signed-off-by:
-
Johannes Berg authored
Signed-off-by:
-
Johannes Berg authored
It's a little confusing to get the BSSID outside the function and pass it in, when it's only needed for this function, so change that. Signed-off-by:
-
Johannes Berg authored
Hopefully the last required fix ... disable beaconing only on beaconing interfaces, and thus avoid calling ieee80211_if_config for purely virtual interfaces (those driver doesn't know about). Signed-off-by:
-
Johannes Berg authored
We cannot currently hand off extra IEs to hw_scan, so reject configuring extra IEs for probe request frames when hw_scan is set. Signed-off-by:
-
Johannes Berg authored
Make sure nobody passes in bogus values, and translate the values (although it isn't necessary). Signed-off-by:
-
Johannes Berg authored
This makes zd1211rw honour enable_beacon, as well as fixing a memory leak: the beacon skb is leaked when configuring it to the hardware fails. Signed-off-by:
-
Johannes Berg authored
Before we have a probe response frame (which is used as the beacon too) there's no need to ask drivers to beacon, they will not get a beacon anyway. Signed-off-by:
-
Johannes Berg authored
It is rather confusing to see this in the log: [ 695.606877] __ratelimit: 40 callbacks suppressed [ 700.726878] __ratelimit: 40 callbacks suppressed without ever seeing any actual message. This is because zd1211rw doesn't compile out the net_ratelimit() call properly if debugging is not enabled. Signed-off-by:
-
Kalle Valo authored
When a directed tim bit is set, mac80211 currently disables power save ands sends a null frame to the AP. But if dynamic power save is disabled, mac80211 will not enable power save ever gain. Fix this by adding ps-poll functionality to mac80211. When a directed tim bit is set, mac80211 sends a ps-poll frame to the AP and checks for the more data bit in the returned data frames. Using ps-poll is slower than waking up with null frame, but it's saves more power in cases where the traffic is low. Userspace can control if either ps-poll or null wakeup method is used by enabling and disabling dynamic power save. Signed-off-by:
Kalle Valo <kalle.valo@nokia.com> Signed-off-by:
-
Kalle Valo authored
Currently mac80211 checks for the multicast tim bit from beacons, disables power save and sends a null frame if the bit is set. This was added to support ath9k. But this is a bit controversial because the AP will send multicast frames immediately after the beacon and the time constraints are really high. Relying mac80211 to be fast enough here might not be reliable in all situations. And there's no need to send a null frame, AP will send the frames immediately after the dtim beacon no matter what. Also if dynamic power save is disabled (iwconfig wlan0 power timeout 0) currently mac80211 disables power save whenever the multicast bit is set but it's never enabled again after receiving the first multicast/broadcast frame. The current implementation is not usable on p54/stlc45xx and the easiest way to fix this is to remove the multicast tim bit check altogether. Handling multicast tim bit in host is rare, most of the designs do this in firmware/hardware, so it's better not to have it in mac80211. It's a lot better to do this in firmware/hardware, or if that's not possible it could be done in the driver. Also renamed the function to ieee80211_check_tim() to follow the style of the file. Signed-off-by:
-
Tobias Klauser authored
The C99 specification states in section 6.11.5: The placement of a storage-class specifier other than at the beginning of the declaration specifiers in a declaration is an obsolescent feature. Signed-off-by:
Tobias Klauser <tklauser@distanz.ch> Signed-off-by:
-
Tobias Klauser authored
The C99 specification states in section 6.11.5: The placement of a storage-class specifier other than at the beginning of the declaration specifiers in a declaration is an obsolescent feature. Signed-off-by:
Jiri Slaby <jirislaby@gmail.com> Signed-off-by:
-
Sujith authored
Signed-off-by:
Sujith <Sujith.Manoharan@atheros.com> Signed-off-by:
-
Sujith authored
This patch removes the various function pointer assignments and unifies them in a single ops structure. Signed-off-by:
-
Sujith authored
This is preparatory work for removing the individual function pointer assignments in eeprom.c Signed-off-by:
-
Sujith authored
Finally, merge these structures and have a single HW specific data structure. Signed-off-by:
-
Sujith authored
A few variables (bssid, bssidmask, curaid) were duplicated in struct ath_softc and in ath_hal, remove them. Signed-off-by:
-
Sujith authored
Signed-off-by:
-
Sujith authored
This patch moves all the HW version/revision specific information into a separate structure. Signed-off-by:
-
Sujith authored
This patch removes the useless sc_ prefixes for all variables. Also, refer to interfaces as VIFs and not as VAPs anymore. Signed-off-by:
-
Sujith authored
This patch converts the ANI channel reference to a pointer, this facilitates moving struct ar5416AniState to ani.h Signed-off-by:
-
Sujith authored
Split the core header files into manageable pieces. Signed-off-by:
-
Nick Kossifidis authored
* We call reset too often and this can result various PHY problems. On config_interface we don't need to reset the whole device. TODO: Create a function hw_fast_reset that only resets the PCU (tx/rx stuff) and not the whole device so that we can use this for stuck beacons etc. Signed-off-by:
Nick Kossifidis <mickflemm@gmail.com> Signed-off-by:
-
Nick Kossifidis authored
* Don't enable BMISS interrupts as we don't handle them yet Signed-off-by:
-
Nick Kossifidis authored
* Update reset and sync with HALs * Clean up eeprom settings and tweaking of initvals and put them on separate functions * Set/Restore 32KHz ref clk operation * Add some more documentation TODO: Spur mitigation, tpc, half/quarter rate, compression etc v2: Address comments from Bob and Felix and fix RSSI threshold bug introduced on the first version of the patch Signed-off-by:
-
Nick Kossifidis authored
* Update initvals to match legacy and Sam's HAL Signed-off-by:
-
Nick Kossifidis authored
* Use the new way to modify rf buffer and put some rf buffer documentation on rfbufer.h * Merge all rf regs functions to one * Sync with legacy HAL and Sam's HAL * Set gain_F settings so that gain_F optimization engine works on RF5111/RF5112 (note that both HALs only use step 0 for RF5111 and they don't use gain_F optimization for this chip, code is there but is never used) Signed-off-by:
-
Nick Kossifidis authored
* Update and cleanup rf gain optimization code * Add comments and refferences to docs and use sane function names * Use only step index on ath5k_gain, no need to have a pointer to the current step since we can determine te step from it's index, this also allows us to put all other structs on rfgain.h and cleanup ath5k.h a little * No need for ah_rfgain variable, we use ah_gain.g_state for everything * Tested on RF2112B chip but gain_F calibration is not yet done (we will finish this on the next patch where we'll rewrite rf-buffer handling) * Use initial rf gain settings for 2316 and 2317 SoCs introduced on a previous patch It seems big but it's mostly cleanup, very few functional changes have been made on phy.c Signed-off-by:
-
Nick Kossifidis authored
* Clean up initial rf buffer settings (new file rfbufer.h) and introduce a new way to access specific rf registers (will use it later) * Clean up initial rf gain settings by moving them on a new file (rfgain.h) so we can later work on gain optimization functions * Update initial rf buffer settings and initial rf gain settings from HALs. This breaks things for now because our current dumps come from pre-configured rf buffer (regdumps already had the needed values set from binary HAL). Signed-off-by:
-
Alina Friedrichsen authored
The atomic requirement for the TSF callbacks is outdated. get_tsf() is only called by ieee80211_rx_bss_info() which is indirectly called by the work queue ieee80211_sta_work(). In the same context are called several other non-atomic functions, too. And the atomic requirement causes problems for drivers of USB wifi cards. Signed-off-by:
Alina Friedrichsen <x-alina@gmx.net> Signed-off-by:
-
