- 09 Dec, 2013 33 commits
-
-
Eyal Shapira authored
Rename for clearer names. Signed-off-by: Eyal Shapira <eyal@wizery.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Expected TPT table was updated only when switching to a new modulation. This is wrong as toggling aggregation changes the expected TPT signficantly. This leads to scenarios where turning aggregation on after being in MIMO sends us back to SISO despite a perfect success ratio. This occurred because the TPT of the SISO mode was being estimated based on aggregation while the MIMO one wasn't. Also remove an error print which isn't an error anymore since we might be updating the expected TPT table due to aggregation changes. Signed-off-by: Eyal Shapira <eyal@wizery.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
This terminates a test window of a new rate quicker in case we've hit a bad rate and should recover better. Signed-off-by: Eyal Shapira <eyal@wizery.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Enable better tracking of different decisions made by the rate scale algorithm. Signed-off-by: Eyal Shapira <eyal@wizery.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Remote peer can publish a different number of supported nss via the operating mode notification IE or action frame. If it limits to 1 then we don't want mimo rates configured in the rate table. Signed-off-by: Eyal Shapira <eyal@wizery.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
MCS9 in NSS=1 or NSS=2 isn't valid for 20Mhz so don't enable it in case we're dealing with a 20Mhz sta. Trying to configure an MCS9 rate with 20Mhz in the LQ rate table would lead to a FW assert. Signed-off-by: Eyal Shapira <eyal@wizery.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Switch to using the new btcoex decision api regarding MIMO and stop accessing the internal btcoex structs. In case MIMO should be disabled it would detect this upon the next Tx and force a search. The search will switch to SISO on a antenna A which isn't used by BT. Signed-off-by: Eyal Shapira <eyal@wizery.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Johannes Berg authored
For now, the firmware doesn't really use the field, but it should be set to zero if there's no specific request. Setting it to the max quota doesn't make sense. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Reviewed-by: Ilan Peer <ilan.peer@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Haim Dreyfuss authored
Advertise driver's support for low priority scan. Notice that this overwrites current setting by mac80211 which depends only on hw scan support. This scan priority can be configured by user space application and it affects scan continuity, low priority scan will be more fragmented scan. Signed-off-by: Haim Dreyfuss <haim.dreyfuss@intel.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eliad Peller authored
Bit 0 in the scan offloading flags asks the filter to pass all the results (instead of filtering them, by default), rather than the other way around (like it is defined and used today). Fix the flag name appropriately, and fix its user. Signed-off-by: Eliad Peller <eliad@wizery.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Emmanuel Grumbach authored
This allows to see the content of the NVM the driver reads. Note that the output is in binary, and requires some external user space tool to display the data properly. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Luciano Coelho authored
If we haven't loaded the D3 image yet and a failure in the suspend process occurs, we shouldn't restart the HW, because we're still running the D0 image. Signed-off-by: Luciano Coelho <luciano.coelho@intel.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Johannes Berg authored
This is just a helper function to go from the mac80211 station struct to our internal one, to later allow us to avoid temporary 'mvmsta' variables. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
valid_rx_ant is a bitmask of available antennas and not the number of Rx antennas. Use num_of_ant and remove duplicate definitions in both dvm and mvm. Signed-off-by: Eyal Shapira <eyal@wizery.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Michal Nazarewicz authored
inta is checked to be zero in a IRQ_NONE branch so afterwards it cannot be zero as it is never modified. Signed-off-by: Michal Nazarewicz <mina86@mina86.com> [reword the patch title and fix comment] Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Emmanuel Grumbach authored
These devices are not sold as discrete modules but are rather soldered down to the board. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Emmanuel Grumbach authored
Latest tests have shown that when BT is active and has connections but the traffic is low, the WiFi aggregation can be large up to 4000us without noticeable impact on BT. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Add more bit field definitions used in LQ flags. Also rename the first bit to a shorter name. Signed-off-by: Eyal Shapira <eyal@wizery.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Johannes Berg authored
Abstract the copy_from_user() pattern into the macros defining debugfs files, reducing the code and making adding new files safer by avoiding having deal with copy_from_user() directly. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Johannes Berg authored
There's really no reason to use void *dbgfs_data rather than a struct iwl_mvm *mvm, so do that. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Johannes Berg authored
The debugfs file is getting pretty large and mixed up between code for the hardware and code for each interface (the two aren't even clearly separated in the code). Make it easier to handle by splitting the per-interface code into a separate file called debugfs-vif.c. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Johannes Berg authored
Just clean up the code a bit, in particular * make all the debugfs writes follow the same pattern with respect to buf/buf_size variables * get rid of useless comments * check return values of all file creations * drop unnecessary parentheses * remove an unused struct definition * fix some whitespace Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Avri Altman authored
The HT protection flags have been a mess for a long time, this patch fixes all this. The proper source of information for this is the protection in the HT operation mode IE which is propagated to bss_conf by mac80211. 1) No need to set the HT protection flags when the link doesn't support HT. 2) We need to set the TGG protection when ERP is active. Signed-off-by: Avri Altman <avri.altman@intel.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eliad Peller authored
Keeping connection can be useful also when testing d3 using the debugfs file (d3_test). Save the vif to keep connection on, and consider it when iterating over the interfaces. Signed-off-by: Eliad Peller <eliad@wizery.com> Reviewed-by: Luciano Coelho <luciano.coelho@intel.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Re-initialize rs when we get a rate_update hook called which happens when remote STA changes bandwidth or sends a VHT opmode notification. Signed-off-by: Eyal Shapira <eyal@wizery.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eliad Peller authored
When waking up from d3 due to disconnection event (e.g. because of beacon loss), we should disconnect immediately. Signed-off-by: Eliad Peller <eliad@wizery.com> Reviewed-by: Luciano Coelho <luciano.coelho@intel.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eliad Peller authored
Even if d3_test_pme_ptr is not available, d3_test can still be useful, e.g. by manually triggering resume (with ctrl-c), and reading the wakeup reasons. Signed-off-by: Eliad Peller <eliad@wizery.com> Reviewed-by: Luciano Coelho <luciano.coelho@intel.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Refactor setting of VHT enabled rates in preparation of adding some more logic there. Signed-off-by: Eyal Shapira <eyal@wizery.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Remove the flags parameter which should be set to sync or async according to whether this is called during sta init or not. Signed-off-by: Eyal Shapira <eyal@wizery.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Eyal Shapira authored
Info about VHT/HT rates wasn't printed properly when dumping the rate scale table. Fix that and print more info. While at it fix some other minor issues in the printing and prevent overflowing the print buffer. Signed-off-by: Eyal Shapira <eyal@wizery.com> Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Johannes Berg authored
After wait_event_timeout(), the condition must still be true if it returns >0, in fact almost the last thing in it is checking the condition again. It's therefore not useful to check yet again in our code, clean it up. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Alexander Bondar authored
Writting to fw_nmi entry will set DEVICE_SET_NMI_REG (0x00a01c30) generating NMI that halts NIC CPU. Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Alexander Bondar <alexander.bondar@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
Emmanuel Grumbach authored
The transport layer doesn't need to know the TX_CMD id. It can be set by the op_mode. The transport layer still needs to know the layout of the Tx command because of alignment issues and because of the scratch pointer. Reviewed-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
-
- 06 Dec, 2013 1 commit
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirelessJohn W. Linville authored
Conflicts: drivers/net/wireless/brcm80211/Kconfig net/mac80211/util.c
-
- 05 Dec, 2013 6 commits
-
-
Luis R. Rodriguez authored
As it stands dynamic user regulatory domain support is only possible for a few programmed regulatory domains as a few countries do not allow for this. The existing code however only would take advantage of the feature if a custom world regulatory domain is used though as that's when we clear beconing flags. We need to lift this restriction as otherwise this feature is pointless. Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
-
Luis R. Rodriguez authored
This will be used later. Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
-
Luis R. Rodriguez authored
This will be used later. Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
-
Luis R. Rodriguez authored
freq_reg_info() expects KHz and not MHz, fix this. In this case we'll now be getting the no-ir flags cleared on channels for any channel when the country IE trusts that channel. @@ struct ieee80211_channel *ch; struct wiphy *wiphy; const struct ieee80211_reg_rule *rule; @@ -rule = freq_reg_info(wiphy, ch->center_freq); +rule = freq_reg_info(wiphy, MHZ_TO_KHZ(ch->center_freq)); Generated-by: Coccinelle SmPL Cc: Julia Lawall <julia.lawall@lip6.fr> Cc: Peter Senna Tschudin <peter.senna@gmail.com> Reported-by: Mihir Shete <smihir@qti.qualcomm.com> Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
-
Luis R. Rodriguez authored
While at it convert this into a switch statement, this makes it easier and manage. Cc: smihir@qti.qualcomm.com Cc: tushnimb@qca.qualcomm.com Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
-
Luis R. Rodriguez authored
Simplify ath_reg_apply_beaconing_flags() by making use of thew new no-ir helper. Cc: smihir@qti.qualcomm.com Cc: tushnimb@qca.qualcomm.com Signed-off-by: Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
-