- 16 Dec, 2011 27 commits
-
-
Don Fry authored
Remove all but the last few references to iwl_priv from the lower level iwl-ucode.c, with resulting code changes. Signed-off-by:
Don Fry <donald.h.fry@intel.com> Signed-off-by:
Wey-Yi Guy <wey-yi.w.guy@intel.com>
-
Don Fry authored
The mac80211 setup_register operations are collected in one routine, but the cleanup routines are not. Create a routine for this. Signed-off-by:
-
Emmanuel Grumbach authored
A few descriptions were missing Signed-off-by:
Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by:
-
Emmanuel Grumbach authored
If the agg SM is in IWL_EMPTYING_HW_QUEUE_ADDBA or in IWL_EMPTYING_HW_QUEUE_DELBA, we are not supposed to get Tx packets from mac80211. mac80211 is supposed to buffer these packets for us. A few issues have been identified in this mechanism, not all of them were fixed. Signed-off-by:
-
Emmanuel Grumbach authored
Since my latest patches, the upper layer reports to mac80211 that the driver is ready to continue the start / stop BA flow as opposed to the transport layer. Hence, iwl_{start,stop}_tx_ba_trans_ready are not needed any more. Signed-off-by: -
Emmanuel Grumbach authored
This will allow us to catch bad cases in which the packets aren't in the right place on the ring. Signed-off-by:
-
Emmanuel Grumbach authored
Since the station is removed, we need to reset the information that was accounted for this station. Signed-off-by:
-
Emmanuel Grumbach authored
The transport doesn't need to access it any more. Signed-off-by: -
Emmanuel Grumbach authored
This is another step towards the move of tid_data from the shared area. Signed-off-by:
-
Emmanuel Grumbach authored
The tid_data is not related to the transport layer, so move the logic that depends on it to the upper layer. This patch deals with the seq_number. Signed-off-by:
-
Emmanuel Grumbach authored
The tid_data is not related to the transport layer, so move the logic that depends on it to the upper layer. This patch deals with the mapping of RA / TID to HW queues in AGG. Signed-off-by:
-
Emmanuel Grumbach authored
The tid_data is not related to the transport layer, so move the logic that depends on it to the upper layer. This patch deals with the code that checks if there are still pending packets for an RA / TID. Signed-off-by:
-
Emmanuel Grumbach authored
The tid_data is not related to the transport layer, so move the logic that depends on it to the upper layer. This patch deals with tx AGG setup. Signed-off-by:
-
Emmanuel Grumbach authored
The tid_data is not related to the transport layer, so move the logic that depends on it to the upper layer. This patch deals with tx AGG alloc. Signed-off-by:
-
Emmanuel Grumbach authored
The tid_data is not related to the transport layer, so move the logic that depends on it to the upper layer. This patch deals with tx AGG stop. Signed-off-by:
-
Emmanuel Grumbach authored
In the same spirit as the previous patch. Eventually this will allow us to remove the tid_data knowledge from the transport layer. Signed-off-by:
-
Emmanuel Grumbach authored
In the very first implementation of HT, the driver was responsible for the queueing: stopping and waking the queues while the HW queues where being drained. In this implementation, we had to deal with the case where we were draining the AGG queue because we wanted to tear down the BA agreement. In the normal flow (when we don't drain any HW queue), when packets are reclaimed, we wake the SW queue in case the SW queue was stopped which can happen when the HW queues are too full. While draining a HW queue, we must make sure that we don't wake the SW queue, since the whole point of the draining is to empty totally the HW queue and not only get below a certain threshold. This is why there is condition in the reclaim function: if (NOT EMPTYING DELBA) wake the SW queue is applicable Since then, a lot has changed and mac80211 is now able to buffer packets that are being sent to a packet list that will be spliced after the driver has reported it has drained its HW queues. Hence, there is no need for the for aforementioned if, and we can safely wake up the queue even if we are draining HW queues. Removing this if, also allows us to remove the wake_queue in check_empty that was there in order to deal with a corner case created by the if. Signed-off-by:
-
Emmanuel Grumbach authored
Since packets sent to an RA / TID in AGG are sent from a separate HW Tx queue, we may get into a race: the regular queue isn't empty while we already begin to send packets from the AGG queue. This would result in sending packets out of order. In order to cope with this, mac80211 waits until the driver reports that the legacy queue is drained before it can send packets to the AGG queue. During that time, mac80211 buffers packets for the driver. These packets will be sent in order after the driver reports it is ready. The way this was implemented in the driver is as follows: We held a counter that monitors the number of packets for an RA / TID in the HW queues. When this counter reached 0, we knew that the HW queues were drained and we reported to mac80211 that were ready to proceed. This patch changes the implementation described above. We now remember what is the wifi sequence number of the first packet that will be sent in the AGG queue (lets' call it ssn). When we reclaim the packet before ssn, we know that the queue is drained, and we are ready to proceed. This will allow us to move this logic in the upper layer and eventually remove the tid_data from the shared area. Signed-off-by:
-
Emmanuel Grumbach authored
ba_resp->seq_ctl is __le16, need to translate to cpu endianity. Signed-off-by:
-
Don Fry authored
Replace the engineering names with the marketing names for the new devices. Signed-off-by:
-
Don Fry authored
Move the configuration pointer from the upper level iwl_priv to the lower level iwl_shared structure, with associated code fixes. Signed-off-by:
-
Chun-Yeow Yeoh authored
Signed-off-by:
Chun-Yeow Yeoh <yeohchunyeow@gmail.com> Acked-by:
Luis R. Rodriguez <mcgrof@qca.qualcomm.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Rajkumar Manoharan authored
When we restore regulatory settings the world regulatory domain is properly reset on cfg80211 (or user prefered regulatory domain) but we were never setting back channel values for drivers that use WIPHY_FLAG_CUSTOM_REGULATORY. Set these values up again by using the orig_ channel parameters. This fixes restoring custom regulatory settings upon disconnect events. Cc: compat@orbit-lab.org Cc: Paul Stewart <pstew@google.com> Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Cc: Senthilkumar Balasubramanian <senthilb@qca.qualcomm.com> Signed-off-by:
Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Signed-off-by:
-
Luis R. Rodriguez authored
By definition WIPHY_FLAG_STRICT_REGULATORY was intended to allow the wiphy to adjust itself to the country IE power information if the card had no regulatory data but we had no way to tell cfg80211 that if the card also had its own custom regulatory domain (these are typically custom world regulatory domains) that we want to follow the country IE's noted values for power for each channel. We add support for this and document it. This is not a critical fix but a performance optimization for cards with custom regulatory domains that associate to an AP with sends out country IEs with a higher EIRP than the one on the custom regulatory domain. In practice the only driver affected right now are the Atheros drivers as they are the only drivers using both WIPHY_FLAG_STRICT_REGULATORY and WIPHY_FLAG_CUSTOM_REGULATORY -- used on cards that have an Atheros world regulatory domain. Cards that have been programmed to follow a country specifically will not follow the country IE power. So although not a stable fix distributions should consider cherry picking this. Cc: compat@orbit-lab.org Cc: Paul Stewart <pstew@google.com> Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Cc: Senthilkumar Balasubramanian <senthilb@qca.qualcomm.com> Reported-by:
-
Johannes Berg authored
Fix a few minor issues with wmediumd_pid: a) make static b) use u32 to match the snd_pid type c) use ACCESS_ONCE since we don't lock it d) don't explicitly initialize to 0 Signed-off-by:
Johannes Berg <johannes.berg@intel.com> Signed-off-by:
-
Mohammed Shafi Shajakhan authored
its not used anywhere in the current code Signed-off-by:
Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com> Signed-off-by:
-
Mohammed Shafi Shajakhan authored
before concluding that the recieved beacon is for us, let us make sure that the BSSID is non-zero. when I configured ad-hoc mode as creator and left it for some time without joining I found we recieved few frames whose BSSID is zero, which we concluded wrongly as 'my_beacons' Signed-off-by:
-
- 15 Dec, 2011 13 commits
-
-
Rafał Miłecki authored
Reported-by:
Rafał Miłecki <zajec5@gmail.com> Signed-off-by:
-
Dan Carpenter authored
Negate has higher precendence than compare and since neither zero nor one are equal to four or eight the original condition is always false. Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Acked-by:
Arend van Spriel <arend@broadcom.com> Signed-off-by:
-
Zefir Kurtisi authored
This initial DFS module provides basic functionality to deal with radar pulses reported by the Atheros DFS HW pulse detector. The reported data is evaluated and basic plausibility checks are performed to filter false pulses. Passing radar pulses are forwarded to pattern detectors which are not yet implemented. (Some modifications to actually use ATH9K_DFS_DEBUGFS based on comments from Julian Calaby <julian.calaby@gmail.com>. -- JWL) Signed-off-by:
Zefir Kurtisi <zefir.kurtisi@neratec.com> Signed-off-by:
-
Zefir Kurtisi authored
In order to enable DFS upstream we want to be sure DFS has been tested for each chipset. Push for public documentation of the requirements we want in place and allow for enabling each chipset through a single upstream commit. Signed-off-by:
-
Zefir Kurtisi authored
This can later be used by other drivers that implement DFS support. Signed-off-by:
-
Larry Finger authored
In ai_attach(), space is allocated for an si_info struct. Immediately after the allocation, routine ai_doattach() is called and that allocated space is set to zero. As no other routine calls ai_doattach(), kzalloc() can be utilized. Signed-off-by:
Larry Finger <Larry.Finger@lwfinger.net> Acked-by:
-
Luis R. Rodriguez authored
A debug level was added to the ath module for printing MCI messages but no documentation was provided. Clarify that MCI is the Message Coexistence Interface, a private protocol used exclusively for WLAN-BT coexistence starting from AR9462. Cc: wtsao@qca.qualcomm.com Cc: rmanohar@qca.qualcomm.com Cc: mohammed@qca.qualcomm.com Cc: senthilb@qca.qualcomm.com Signed-off-by:
-
Felix Fietkau authored
The receiver will discard them anyway. Signed-off-by:
Felix Fietkau <nbd@openwrt.org> Signed-off-by:
-
Felix Fietkau authored
Instead of releasing and taking back the lock over and over again in the tx path, hold the lock a bit longer, requiring much fewer lock/unlock pairs. This makes locking much easier to review and should not have any noticeable performance/latency impact. Signed-off-by:
-
Felix Fietkau authored
tid->seq_next is initialized on A-MPDU start anyway, and the comment next to this chunk of code seems to be bogus as well. Signed-off-by:
-
Felix Fietkau authored
Signed-off-by:
-
Felix Fietkau authored
When processing A-MPDU tx status, only send a BAR for the failed packet with the highest sequence number. Signed-off-by:
-
Felix Fietkau authored
Instead of limiting a subframe to 10 A-MPDU software transmission attempts, count hardware retransmissions as well and raise the limit a bit. That way there will be fewer software retransmission attempts when traffic suffers from lots of hardware retransmissions. Signed-off-by:
-
