- 15 Jul, 2020 19 commits
-
-
Tony Lindgren authored
We currently have a collection of flags and locking between the threaded irq and tx work: - wl->flags bitops - wl->mutex - wl->wl_lock spinlock The bitops flags do not need a spinlock around them, and we only need the spinlock to see if we need to queue tx work or not. And wlcore_irq() holds the mutex. To simplify the locking, we can use spin_trylock and always queue tx work unless we know there's nothing to do. Let's also update the comment a bit while at it. Signed-off-by:
Tony Lindgren <tony@atomide.com> Signed-off-by:
Kalle Valo <kvalo@codeaurora.org> Link: https://lore.kernel.org/r/20200702162951.45392-4-tony@atomide.com
-
Tony Lindgren authored
We currently have a collection of flags and locking between the threaded irq and tx work: - wl->flags bitops - wl->mutex - wl->wl_lock spinlock The bitops flags do not need a spinlock around them, and wlcore_irq() already holds the mutex calling wlcore_irq_locked(). And we only need the spinlock to see if we need to run the queue or not. To simplify the locking, we can use spin_trylock and always run the tx queue unless we know there's nothing to do. Signed-off-by:
-
Tony Lindgren authored
We can simplify the runtime resume ELP path by always setting and clearing the completion in runtime resume. This way we can test for WL1271_FLAG_IRQ_RUNNING after the resume write to see if we need completion at all. And in wlcore_irq(), we need to take spinlock for running the completion and for the pm_wakeup_event(). Spinlock is not needed around the bitops flags check for WL1271_FLAG_SUSPENDED so the spinlocked sections get shorter. Signed-off-by:
-
Pali Rohár authored
This macro is already used in mwifiex driver for specifying upper limit and is defined to value 3. So use it also in struct ieee80211_iface_limit. Signed-off-by:
Pali Rohár <pali@kernel.org> Signed-off-by:
-
Yan-Hsuan Chuang authored
A new RFE type 6 of RTL8822CE is released, add it in the RFE type definition. Signed-off-by:
Yan-Hsuan Chuang <yhchuang@realtek.com> Signed-off-by:
-
Aaron Ma authored
New device ID 0xc82f found on Lenovo ThinkCenter. Tested it with c822 driver, works good. PCI id: 03:00.0 Network controller [0280]: Realtek Semiconductor Co., Ltd. Device [10ec:c82f] Subsystem: Lenovo Device [17aa:c02f] Signed-off-by:Aaron Ma <aaron.ma@canonical.com> Signed-off-by:
-
Yan-Hsuan Chuang authored
Some platforms cannot read the DBI register successfully for the ASPM settings. After the read failed, the bus could be unstable, and the device just became unavailable [1]. For those platforms, the ASPM should be disabled. But as the ASPM can help the driver to save the power consumption in power save mode, the ASPM is still needed. So, add a module parameter for them to disable it, then the device can still work, while others can benefit from the less power consumption that brings by ASPM enabled. [1] https://bugzilla.kernel.org/show_bug.cgi?id=206411 [2] Note that my lenovo T430 is the same. Fixes: 3dff7c6e ("rtw88: allows to enable/disable HCI link PS mechanism") Signed-off-by:
-
Tzu-En Huang authored
Since 8821C code is ready, we can build it. Signed-off-by:
Tzu-En Huang <tehuang@realtek.com> Signed-off-by:
-
Tzu-En Huang authored
Since single rf path chips don't support TX SPBC, tell mac80211 to not advertise it. Signed-off-by:
-
Tzu-En Huang authored
Beamforming is used for directional signal transmission/reception. Beamformee plays the role for signal reception, and makes the RX performance better in middle distance range. Implement beamformee related callbacks for 8821c. Since 8821c only support 1ss rate, nc_index in beamformee setting needs to be adjusted based on the capability. Signed-off-by:
-
Tzu-En Huang authored
The TX power requires to be adjusted based on the thermal value. The actual power will decrease if the thermal value raised, and will increase if the thermal value lowered. Driver comapres the thermal value, as moving averages. If it changes over a limit, driver will modify the TX power index to compensate. Implement rtw_chip_ops::pwr_track() for 8821c. Signed-off-by:
-
Tzu-En Huang authored
CCK PD can reduce the number of false alarm of the CCK rates. It dynamically adjusts the power threshold and CS ratio. The values are compared to the values of the previous level, if the level is changed, set new values of power threshold and CS ratio. Implement rtw_chip_ops::cck_pd_set() for 8821c. Signed-off-by:
-
Tzu-En Huang authored
In order to get a better TX EVM, do calibration after association. The calibration needed for 8821c is the IQK, which is done in the firmware. Implement the rtw_chip_ops::phy_calibration() to trigger firmware to calibrate. Reviewed-by:
Sebastian Andrzej Siewior <bigeasy@linutronix.de> Signed-off-by:
-
Tzu-En Huang authored
False alarm statistics can be used to adjust the RX gain. This helps the driver to adapt to different circumstances. Implement rtw_chip_ops::false_alarm_statistics() for 8821c. Reviewed-by:
-
Tzu-En Huang authored
Some RX packets contain also information about environment status. Implement rtw_chip_ops::query_rx_desc() for 8821c. Parse the RX descriptor which describes the current condition of the received packet. Reviewed-by:
-
Tzu-En Huang authored
8821c is capable of 2.4G and 5G. Implement rtw_chip_ops::set_channel() to set 2G and 5G channels. This includes MAC, BB and RF related settings. Reviewed-by:
-
Tzu-En Huang authored
To improve user experience in field, we need DIG to adjust RX initial gain depends on field situation. Define the register addresses for 8821c. Reviewed-by:
-
Tzu-En Huang authored
To configure the transmit power of 8821c implement trasmit power index setting callback function for 8821c. This is very similar to the callback function of 8822b. Reviewed-by:
-
Tzu-En Huang authored
RTL8821CE chipsets are 802.11ac dual-band WiFi + BT combo chips. This patch adds the basic functions such as parameter tables, chip information, power on flow. Reviewed-by:
-
- 14 Jul, 2020 21 commits
-
-
Luc Van Oostenryck authored
The method ndo_start_xmit() is defined as returning an 'netdev_tx_t', which is a typedef for an enum type defining 'NETDEV_TX_OK' but this driver returns '0' instead of 'NETDEV_TX_OK'. Fix this by returning 'NETDEV_TX_OK' instead of '0'. Signed-off-by:
Luc Van Oostenryck <luc.vanoostenryck@gmail.com> Signed-off-by:
-
Ajay Singh authored
Added version number info along with firmware name so driver can pick the correct revision of FW file. Moved FW filename macro as part of driver code & added MODULE_FIRMWARE to specify FW needed by module. Signed-off-by:
Ajay Singh <ajay.kathat@microchip.com> Signed-off-by:
-
Ajay Singh authored
Modify WILC1000 binary filename to use single unified wilc1000 FW. A single wilc1000 binary is used for different wilc1000 revisions. Signed-off-by:
-
Ajay Singh authored
Avoid below reported warning found when 'CONFIG_PM' config is undefined. 'warning: unused variable 'wowlan_support' [-Wunused-const-variable]' Reported-by:
kernel test robot <lkp@intel.com> Signed-off-by:
-
Ajay Singh authored
Make use 'strlcpy' instead of 'strncpy' to overcome 'stringop-truncation' compiler warning. Reported-by:
-
Nicolas Ferre authored
As Adham's email address is bouncing, remove him from wilc1000 entry and add Claudiu as a new co-maintainer. Claudiu follows wilc1000 driver development for a long time and contributed to it already. Signed-off-by:
Nicolas Ferre <nicolas.ferre@microchip.com> Signed-off-by:
-
Linus Walleij authored
The code in the BRCM80211 BRCMSMAC driver is using the legacy GPIO API to to a complex check of the validity of the base of the GPIO chip and whether it is present at all and then adding an offset to the base of the chip. Use the existing function to obtain a GPIO line internally from a GPIO chip so we can use the offset directly and modernize the code to use GPIO descriptors instead of integers from the global GPIO numberspace. Cc: Wright Feng <wright.feng@cypress.com> Cc: Frank Kao <frank.kao@cypress.com> Cc: Kalle Valo <kvalo@codeaurora.org> Signed-off-by:
Linus Walleij <linus.walleij@linaro.org> Signed-off-by:
-
Matthias Brugger authored
The driver relies on the compatible string from DT to determine which FW configuration file it should load. The DTS spec allows for '/' as part of the compatible string. We change this to '-' so that we will still be able to load the config file, even when the compatible has a '/'. This fixes explicitly the firmware loading for "solidrun,cubox-i/q". Signed-off-by:
Matthias Brugger <mbrugger@suse.com> Reviewed-by:
Hans deGoede <hdegoede@redhat.com> Signed-off-by:
-
Wright Feng authored
Linux 3.6 introduces TSQ which has a per socket threshold for TCP Tx packet to reduce latency. In flow control mode, host driver enqueues skb in hanger and TCP doesn't push new skb frees until host frees the skb when receiving fwstatus event. So set pacing shift 8 to send them as a single large aggregate frame to the bus layer. 43455 TX TCP throughput in different FC modes on Linux 5.4.18 sk_pacing_shift : Throughput (fcmode=0) 10: 245 Mbps 9: 245 Mbps 8: 246 Mbps 7: 246 Mbps sk_pacing_shift : Throughput (fcmode=1) 10: 182 Mbps 9: 197 Mbps 8: 206 Mbps 7: 207 Mbps sk_pacing_shift : Throughput (fcmode=2) 10: 180 Mbps 9: 197 Mbps 8: 206 Mbps 7: 207 Mbps Signed-off-by:
Wright Feng <wright.feng@cypress.com> Signed-off-by:
Chi-hsien Lin <chi-hsien.lin@cypress.com> Signed-off-by:
-
Wright Feng authored
When USB or SDIO device got abnormal bus disconnection, host driver tried to clean up the skbs in PSQ and TXQ (The skb's pointer in hanger slot linked to PSQ and TSQ), so we should set the state of skb hanger slot to BRCMF_FWS_HANGER_ITEM_STATE_FREE before freeing skb. In brcmf_fws_bus_txq_cleanup it already sets BRCMF_FWS_HANGER_ITEM_STATE_FREE before freeing skb, therefore we add the same thing in brcmf_fws_psq_flush to avoid following warning message. [ 1580.012880] ------------ [ cut here ]------------ [ 1580.017550] WARNING: CPU: 3 PID: 3065 at drivers/net/wireless/broadcom/brcm80211/brcmutil/utils.c:49 brcmu_pkt_buf_free_skb+0x21/0x30 [brcmutil] [ 1580.184017] Call Trace: [ 1580.186514] brcmf_fws_cleanup+0x14e/0x190 [brcmfmac] [ 1580.191594] brcmf_fws_del_interface+0x70/0x90 [brcmfmac] [ 1580.197029] brcmf_proto_bcdc_del_if+0xe/0x10 [brcmfmac] [ 1580.202418] brcmf_remove_interface+0x69/0x190 [brcmfmac] [ 1580.207888] brcmf_detach+0x90/0xe0 [brcmfmac] [ 1580.212385] brcmf_usb_disconnect+0x76/0xb0 [brcmfmac] [ 1580.217557] usb_unbind_interface+0x72/0x260 [ 1580.221857] device_release_driver_internal+0x141/0x200 [ 1580.227152] device_release_driver+0x12/0x20 [ 1580.231460] bus_remove_device+0xfd/0x170 [ 1580.235504] device_del+0x1d9/0x300 [ 1580.239041] usb_disable_device+0x9e/0x270 [ 1580.243160] usb_disconnect+0x94/0x270 [ 1580.246980] hub_event+0x76d/0x13b0 [ 1580.250499] process_one_work+0x144/0x360 [ 1580.254564] worker_thread+0x4d/0x3c0 [ 1580.258247] kthread+0x109/0x140 [ 1580.261515] ? rescuer_thread+0x340/0x340 [ 1580.265543] ? kthread_park+0x60/0x60 [ 1580.269237] ? SyS_exit_group+0x14/0x20 [ 1580.273118] ret_from_fork+0x25/0x30 [ 1580.300446] ------------ [ cut here ]------------ Acked-by:
Arend van Spriel <arend.vanspriel@broadcom.com> Signed-off-by:
-
Chi-Hsien Lin authored
commit 4684997d ("brcmfmac: reset PCIe bus on a firmware crash") adds a reset function to recover firmware trap for PCIe bus. This commit adds an implementation for SDIO bus. Upon SDIO firmware trap, do below: - Remove the device - Reset hardware - Probe the device again Signed-off-by:
-
Able Liao authored
Ignore FW event if the event's BSSID is different form the BSSID of the currently connected AP. Check interface state is connected or not, if state is not connected that can ignore link down event. Signed-off-by:
Able Liao <Able.Liao@cypress.com> Signed-off-by:
-
Soontak Lee authored
Current brcmf_link_down() always call cfg80211_disconnected() with locally_generated=1, which is not always the case. Add event source argument on link down handler and set locally_generated based on the real trigger. Signed-off-by:
Soontak Lee <soontak.lee@cypress.com> Signed-off-by:
-
Soontak Lee authored
Unable to change back to visiable SSID because there is no disable hidden ssid routine. Signed-off-by:
-
Joseph Chuang authored
Commit 4905432b28b7 ("brcmfmac: Fix P2P Group Formation failure via Go-neg method") did not initialize requested_dwell properly, resulting in an always-false dwell time overflow check. Fix it by setting the correct requested_dwell value. Fixes: 4905432b28b7 ("brcmfmac: Fix P2P Group Formation failure via Go-neg method") Reported-by:Joseph Chuang <joseph.chuang@cypress.com> Signed-off-by:
-
Double Lo authored
This patch move the credit map setting to right place to avoid brcmf_fws_return_credits() return without setting the credit map. It fix the thoughput zero stalls issue in softAP mode when STA using PM 1 mode. Signed-off-by:
Double Lo <double.lo@cypress.com> Signed-off-by:
-
Chung-Hsien Hsu authored
There is a mismatch of tx status flag values between host and firmware. It makes the host mistake the flags and have incorrect behavior of credit returns. So update the flags to sync with the firmware ones. Signed-off-by:
Chung-Hsien Hsu <stanley.hsu@cypress.com> Signed-off-by:
-
Amar Shankar authored
It is observed that sometimes when sdiod is low in tx credits in low rssi scenarios, the data path consumes all sdiod rx all credits and there is no sdiod rx credit available for control path causing host and card to go out of sync resulting in link loss between host and card. So in order to prevent it some credits are reserved for control path. Note that TXCTL_CREDITS can't be larger than the firmware default credit update threshold 2; otherwise there will be a deadlock for both side waiting for each other. Signed-off-by:
Amar Shankar <amsr@cypress.com> Signed-off-by:
-
Jia-Shyr Chuang authored
In wifi firmware, max length of IOCTL/IOVAR buffer size is 8192. Increase the message buffer max size same as wifi firmware for control packets so return buffers can come back. Signed-off-by:
-
Raveendran Somu authored
Current credit borrowing allows only the access category BE to borrow the credits. This change is to fix the credit borrowing logic, to make borrowing available for all access categories and also to borrow only from the lower categories. This fixes WFA 802.11n certs 5.2.27 failures. Signed-off-by:
Raveendran Somu <raveendran.somu@cypress.com> Signed-off-by:
-
Prasanna Kerekoppa authored
Bss info flag definition need to be fixed from 0x2 to 0x4 This flag is for rssi info received on channel. All Firmware branches defined as 0x4 and this is bug in brcmfmac. Signed-off-by:
Prasanna Kerekoppa <prasanna.kerekoppa@cypress.com> Signed-off-by:
-
