- 24 Sep, 2012 40 commits
-
-
Arend van Spriel authored
Small step to fix structure commenting using kernel-doc syntax. Reviewed-by:
Hante Meuleman <meuleman@broadcom.com> Signed-off-by:
Arend van Spriel <arend@broadcom.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Hante Meuleman authored
On a usb error the lock bit should be cleared. Reviewed-by:
-
Piotr Haber authored
This patch fixes a bug when device is being started while RfKill switch is engaged, leading to hang due to partial initialization of hardware. Tested-by: <dragonn@op.pl> Reviewed-by:
Piotr Haber <phaber@broadcom.com> Signed-off-by:
-
Hante Meuleman authored
clean up code. Reviewed-by:
-
Hante Meuleman authored
This patch fixes a bug in routine where phy bands are determined. Reviewed-by:
-
Hante Meuleman authored
clean up code. Reviewed-by:
-
Franky Lin authored
SDIO read non-glomming frame routine handles first frame and follow up frame read separately. But they share a lot of common code. This patch abstracts a brcmf_sdio_hdparser function and optimize the code flow for better readability and future optimization. Reviewed-by:
Franky Lin <frankyl@broadcom.com> Signed-off-by:
-
Hante Meuleman authored
For suspend/resume it is necessary to store firmware in memory. In order to support multiple usb dongles at the same time a linked list of firmwares was created. Reviewed-by:
Pieter-Paul Giesberts <pieterpg@broadcom.com> Signed-off-by:
-
Hante Meuleman authored
Some events result in printing of the buffer when debug is enabled. This printing was not very efficient. Changed to macro so it comes out nice and clean without filling log buffer. Reviewed-by:
-
Arend van Spriel authored
This change add support for NL80211 scheduled scan. This may be used to offload scanning to the device, which may give the host opportunity to sleep. The newer versions of wpa_supplicant have support for this functionality. Reviewed-by:
-
Arend van Spriel authored
With the introduction of E-Scan there are two scan mechanisms in the driver. I-Scan was aborted on suspend and bringing down the device using brcmf_term_iscan(). The function has been renamed to brcmf_abort_scanning() and covers e-scan abort as well. Reviewed-by:
-
Arend van Spriel authored
With the introduction of e-scan mechanism in brcmfmac (e756af5b brcmfmac: add e-scan support.) a couple of sparse warnings were introduced. This patch resolves those. Reported-by:
Fengguang Wu <fengguang.wu@intel.com> Reviewed-by:
-
Amitkumar Karwar authored
When delayed scan operation is resumed, we just add next scan command in queue but don't wakeup main thread to process the command. Hence the command is downloaded to firmware only after waking up the main thread by any other means. This bug which was introduced after "mwifiex: improve scan delay logic.." patch is fixed here. Signed-off-by:
Amitkumar Karwar <akarwar@marvell.com> Signed-off-by:
Bing Zhao <bzhao@marvell.com> Signed-off-by:
-
Gertjan van Wingerde authored
Setting of the individual fields of the RF register can be simplified. Signed-off-by:
Gertjan van Wingerde <gwingerde@gmail.com> Acked-by:
Ivo van Doorn <IvDoorn@gmail.com> Signed-off-by:
-
Gertjan van Wingerde authored
Signed-off-by:
-
Gertjan van Wingerde authored
Move RT3290 BBP initialization sequence (part) to the right place in the code. This is just a code style change, no functional changes. Signed-off-by:
-
Gertjan van Wingerde authored
Align with the values used by the RT3572 Ralink vendor driver v2.5.0.0. Signed-off-by:
-
Gertjan van Wingerde authored
Make the code in rt2800_get_default_vgc more understandable and readable, especially for the 5GHz band values. Signed-off-by:
-
Sujith Manoharan authored
Signed-off-by:
Sujith Manoharan <c_manoha@qca.qualcomm.com> Signed-off-by:
-
Sujith Manoharan authored
Signed-off-by:
-
Sujith Manoharan authored
This patch adds a new callback to handle WLAN RX diversity for AR9565. Signed-off-by:
-
Sujith Manoharan authored
Signed-off-by:
-
Sujith Manoharan authored
Signed-off-by:
-
Sujith Manoharan authored
Signed-off-by:
-
Sujith Manoharan authored
Set main_gaintb and alt_gaintb once instead of zero'ing them in every case. Signed-off-by:
-
Sujith Manoharan authored
Instead of trying to determine whether antenna diversity is enabled by parsing eeprom data, use ATH9K_HW_CAP_ANT_DIV_COMB which is already filled at init time. Signed-off-by:
-
Mohammed Shafi Shajakhan authored
Update the 'hw_timer_enabled' to 'false' wherever we are stopping hardware generic timers, excecpt the case where we start them again immediately. Signed-off-by:
Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com> Signed-off-by:
-
Arend van Spriel authored
Upon handling an e-scan result event from the firmware in brcmf_cfg80211_escan_handler() the status may indicate e-scan is completed or aborted. If so brcmf_notify_escan_complete() is called. However, that function and brcmf_cfg80211_escan_handler() both log the same debug message and both enable MPC flag in the firmware. Reviewed-by:
-
Arend van Spriel authored
This patch assures data is only appended in output buffer when data is not NULL and datalen is not zero. Reviewed-by:
-
Hante Meuleman authored
This patch fixes bug where driver frees resources twice when an exception occurs. Reviewed-by:
-
Franky Lin authored
SDIO in-band interrupt is level sensitive according to SDIO standard. When the register interrupt handler gets called by SDIO stack it is running in non interrupt context and expected to clear the interrupt from the dongle. Therefore in-band and out-of-band interrupt need to be handled differently. Cc: Wei Ni <wni@nvidia.com> Cc: Stephen Warren <swarren@wwwdotorg.org> Reported-by:
Wei Ni <wni@nvidia.com> Reviewed-by:
Stephen Warren <swarren@wwwdotorg.org> Signed-off-by:
-
Franky Lin authored
Use SDIO host lock to protect the entire communication process of register access or data read/write rather than function calls to MMC/SDIO stack only. This can help to avoid unintentional memory access and unexpected interruption. Reviewed-by:
-
Franky Lin authored
Streamline SDIO dpc by removing some unnecessary code path. Reviewed-by:
-
Franky Lin authored
Switch SDIO dpc implementation from kernel thread to workqueue for better performance and compatibility. Reviewed-by:
-
Franky Lin authored
Interrupt pending flag used in SDIO bus layer could be used in multiple processes in different context. Use atomic_t make sure every interrupt is handled. Reviewed-by:
-
Franky Lin authored
Remove sdio bus sleep mechanism since it is never invoked. Reviewed-by:
-
Franky Lin authored
brcmf_sendpkt only called by brcmf_netdev_start_xmit now. Absorb it to increase readability. Reviewed-by:
-
Ming Lei authored
Inside ath9k_hif_usb_firmware_fail(), the instance of 'struct struct hif_device_usb' may be freed by ath9k_hif_usb_disconnect() after complete(&hif_dev->fw_done); But 'hif_dev' is still accessed after the line code above is executed. This patch fixes the issue by not accessing 'hif_dev' after 'complete(&hif_dev->fw_done)' inside ath9k_hif_usb_firmware_fail(). Cc: ath9k-devel@lists.ath9k.org Cc: "Luis R. Rodriguez" <mcgrof@qca.qualcomm.com> Cc: Jouni Malinen <jouni@qca.qualcomm.com> Cc: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com> Cc: Senthil Balasubramanian <senthilb@qca.qualcomm.com> Cc: "John W. Linville" <linville@tuxdriver.com> Signed-off-by:
Ming Lei <ming.lei@canonical.com> Signed-off-by:
-
Rajkumar Manoharan authored
Right now MCI work is being queued twice by profile and status updation. Queue MCI work once when it is needed. Signed-off-by:
Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Signed-off-by:
-
Rajkumar Manoharan authored
Ensure that led pin is filled and set to OFF before starting the driver. With recent changes, drv_start is being called even before led_init is being completed. This is causing led is always OFF on driver load when the interface is UP. This patch splits the led init and fills the led pin before register hw. Signed-off-by:
-
