- 12 Jun, 2013 40 commits
-
-
Vladimir Kondratiev authored
No more need for special processing of EAPOL, FW can now send EAPOL frames using normal Tx queue for TID 0 This fixes "schedule while atomic" bug - start_xmit called in softirq context; while WMI mechanism that was used may sleep. Signed-off-by:
Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Vladimir Kondratiev authored
When starting secure AP, in some cases wpa_s provides probe template but not probe/assoc IE's. In this case, derive missing IE's from probe. Signed-off-by:
-
Vladimir Kondratiev authored
Report FC from the frame itself, as auxiliary information includes only frame subtype. This is preparation for future changes, when DMG beacon (extension frame) may be reported through wmi_evt_rx_mgmt() Signed-off-by:
-
Vladimir Kondratiev authored
map card's back-door debug data Signed-off-by:
-
Vladimir Kondratiev authored
It may take up to 3500ms for the FW to start AP/PCP. Increase accordingly, adding some safety margin. Signed-off-by:
-
Felix Fietkau authored
The CAB (Content after Beacon) queue is used for beacon-triggered transmission of buffered multicast frames. If lots of multicast frames were buffered and this queue fills up, it drowns out all regular traffic. To limit the damage that buffered traffic can do, try to limit the queued data to becaon_interval / 8. Signed-off-by:
Felix Fietkau <nbd@openwrt.org> Signed-off-by:
-
Felix Fietkau authored
This adds support for PS-Poll and U-APSD driver-buffered frames (part of an aggregation session). Signed-off-by:
-
Felix Fietkau authored
Use the UAPSD hardware queue to get PS-Poll responses out as fast as possible and without backoff. Signed-off-by:
-
Arend van Spriel authored
When initialization within brcmf_bus_start() fails on steps before the brcmf_net_attach() the net_device for the primary interface needs to be freed. This patch resolves a panic during kernel boot as reported by Stephen Warren. ref.: http://mid.gmane.org/51AD1F22.2080004@wwwdotorg.orgTested-by:
Stephen Warren <swarren@nvidia.com> Reviewed-by:
Hante Meuleman <meuleman@broadcom.com> Reviewed-by:
Pieter-Paul Giesberts <pieterpg@broadcom.com> Signed-off-by:
Arend van Spriel <arend@broadcom.com> Signed-off-by:
-
Arend van Spriel authored
When data is available and fw signalling is enabled then TIM information should be sent to firmware. If it can piggy back on existing packet then do that otherwise create dummy packet to get information out. Cc: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by:
-
Arend van Spriel authored
In brcmf_net_attach() the hard_header_len is set to sum of ETH_HLEN and the headroom needed by the bus interface. Better use increment instead as hard_header_len is already initialized upon alloc_netdev(). Reviewed-by:
-
Arend van Spriel authored
The BDC header contains PropTx TLV signals that are useful to capture for debugging. This event captures the header and tlv's in binary form. This can be post-processed using trace-cmd plugin. Reviewed-by:
Franky (Zhenhui) Lin <frankyl@broadcom.com> Signed-off-by:
-
Arend van Spriel authored
The statistic counter send_pkts was wrongly counted conditionally. Correcting the mistake. Reviewed-by:
-
Hante Meuleman authored
Reviewed-by:
-
Hante Meuleman authored
Reviewed-by:
-
Hante Meuleman authored
Currently suppressed packets get enque-ed with header which then gets pulled before transmit. It is more logical and clean to pull the header on return and push it unconditionally on xmit. Reviewed-by:
-
Hante Meuleman authored
During P2P testing it turned out that the firmware sents multiple multiple creditmap event messages. Only the first message from the firmware should be processed. Otherwise the firmware-signalled flow control can run haywire when it has packets outstanding in firmware. Reviewed-by:
-
Hante Meuleman authored
The destination entries for firmware-signalled flow control have the interface id stored. This needs to be translated to bsscfg index when looking up the ifp object for the interface. Reviewed-by:
-
Hante Meuleman authored
In case of TDLS find the correct MAC descriptor for fw signalling data. In case of TDLS each destination gets its own entry. This was not handled correctly for P2P client. Reviewed-by:
-
Hante Meuleman authored
When credit borrowing is active the BE credits have been depleted, however the worker should still be scheduled. In case of credit borrowing correct credit map to make sure worker remains active. Reviewed-by:
-
Hante Meuleman authored
Store gen bit on suppressed packet per entry and use latest stored version for each packet which gets transmitted to fw. Reviewed-by:
-
Hante Meuleman authored
Currently on a bus flow control both fws is informed and netif queue gets closed. In case of fw signalling enabled, let the flow control be handled by fw signalling only. Reviewed-by:
-
Arend van Spriel authored
The debug logging in firmware-signalling code was rather extensive and for a large part in the data path. This patch removes large part or the level is changed to DATA level. Reviewed-by:
-
Arend van Spriel authored
Firmware can request the driver for transmit packets using two different signals. Only for one signal a flag was set in the sk_buff control buffer. This patch adds explicit flag for the other signal as well. Reviewed-by:
-
Arend van Spriel authored
Reworked brcmf_skb_pick_up_credit() so it can be used for both fcmode flavours in the same way. Reviewed-by:
-
Hante Meuleman authored
On bus flow control (no more host bus resources to send packets to device) the netif flow control was toggled, however credit management should also take this status into account. Since there are multiple sources handling this flow control necessary spinlocks were added to protect flow control related data/states. Reviewed-by:
-
Arend van Spriel authored
The brcmf_skbuff_cb structure contain if_flags and htod fields. Both have a bitfield defined to hold the fifo number. With a small code change we get rid of the fifo bitfield in if_flags. Reviewed-by:
-
Arend van Spriel authored
The firmware-signal API specification defines length for the different tlv. During testing on different devices it turned out not all firmware used the tlv length according specification. Therefore the length check is made less strict with this patch. Reviewed-by:
-
Gabor Juhos authored
It is exactly the same like the generic rt2800usb_ops. Remove the duplicate and use the generic ops for all devices. Signed-off-by:
Gabor Juhos <juhosg@openwrt.org> Acked-by:
Gertjan van Wingerde <gwingerde@gmail.com> Signed-off-by:
-
Gabor Juhos authored
The extra_tx_headroom field of struct rt2x00_ops indicates the extra TX headroom size required for a given device. This data is redundant, the value can be computed from the desc_size and winfo_size fields of the TX queues. Move the extra_tx_headroom field to struct rt2x00_dev, compute its value in the probe routine and use the cached value in the rest of the code. Signed-off-by:
Stanislaw Gruszka <sgruszka@redhat.com> Acked-by:
-
Luis R. Rodriguez authored
This adds VHT80 support for the QCA world regulatory domains. Cc: kvalo@qca.qualcomm.com Signed-off-by:
Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by:
-
Wei Yongjun authored
Fix to return a negative error code in the il3945_hw_set_hw_params() error handling case instead of 0, as done elsewhere in this function. Signed-off-by:
Wei Yongjun <yongjun_wei@trendmicro.com.cn> Acked-by:
-
Avinash Patil authored
This patch disables tx_amsdu support in mwifiex by default. tx_amdsu support can be enabled via module parameter at load time. Signed-off-by:
Avinash Patil <patila@marvell.com> Signed-off-by:
Bing Zhao <bzhao@marvell.com> Signed-off-by:
-
Amitkumar Karwar authored
PATCH "mwifiex: scan delay timer cleanup in unload path" adds code to cancel scan delay timer in unload path. It causes a regression for USB interface. USB8797 card gets enumerated twice. First enumeration is for firmware download and second enumeration expects firmware initialization. It was observed that we are trying del_timer_sync() without setting up the timer when remove handler is called after first enumeration. This patch moves setup_timer() call to appropriate place so that timer is setup for both the enumerations. Signed-off-by:
Amitkumar Karwar <akarwar@marvell.com> Signed-off-by:
-
Gabor Juhos authored
If the queue_init callback is implemented by a driver it gets used instead of the data_queue_desc based initialization. The queue_init callback is implemented for each drivers now, so the old initialization method is not used anymore. Remove the unused data_queue_desc structure and all of the related code. Signed-off-by:
-
Gabor Juhos authored
The generic rt2x00 code has been changed to allow the drivers toimplement dynamic data_queue initialization. Remove the static data queue descriptor structures and implement the queue_init callback instead. Compile tested only. Signed-off-by:
-
Gabor Juhos authored
The generic rt2x00 code has been changed to allow the drivers toimplement dynamic data_queue initialization. Remove the static data queue descriptor structures and implement the queue_init callback instead. Signed-off-by:
Jakub Kicinski <kubakici@wp.pl> Acked-by:
-
Gabor Juhos authored
The generic rt2x00 code has been changed to allow the drivers toimplement dynamic data_queue initialization. Remove the static data queue descriptor structures and implement the queue_init callback instead. Compile tested only. Signed-off-by:
-
Gabor Juhos authored
The generic rt2x00 code has been changed to allow the drivers toimplement dynamic data_queue initialization. Remove the static data queue descriptor structures and implement the queue_init callback instead. Compile tested only. Signed-off-by:
-
Gabor Juhos authored
The generic rt2x00 code has been changed to allow the drivers toimplement dynamic data_queue initialization. Remove the static data queue descriptor structures and implement the queue_init callback instead. Signed-off-by:
-
