- 16 Dec, 2021 3 commits
-
-
David Mosberger-Tang authored
The wakeup and sleep commands need to be sent to the WILC chip only when it is in power save mode (PSM, as controlled by "iw dev wlan0 set power_save on/off"). The commands are relatively costly, so it pays to skip them when possible. iperf3 without this patch (no significant different with PSM on/off): TX 0.00-120.01 sec 140 MBytes 9.82 Mbits/sec RX 0.00-120.69 sec 283 MBytes 19.6 Mbits/sec with this patch applied: PSM off (TX is 46% improved, RX slightly improved; may not be significant): TX 0.00-120.00 sec 206 MBytes 14.4 Mbits/sec RX 0.00-120.48 sec 322 MBytes 22.4 Mbits/sec PSM on (no significant change): TX 0.00-120.00 sec 140 MBytes 9.78 Mbits/sec RX 0.00-120.08 sec 257 MBytes 18.0 Mbits/sec Signed-off-by:
David Mosberger-Tang <davidm@egauge.net> Signed-off-by:
Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20211210203016.3680425-2-davidm@egauge.net
-
Merlijn Wajer authored
This fix is similar to commit 77c91295 ("wil6210: specify max. IE length"). Without the max IE length set, wpa_supplicant cannot operate using the nl80211 interface. This patch is a workaround - the number 512 is taken from the wlcore driver, but note that per Paul Fertser: there's no correct number because the driver will ignore the data passed in extra IEs. Suggested-by:
Paul Fertser <fercerpav@gmail.com> Signed-off-by:
Merlijn Wajer <merlijn@wizzup.org> Signed-off-by:
-
Kalle Valo authored
Merge tag 'iwlwifi-next-for-kalle-2021-12-08' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-next first set of iwlwifi patches for v5.17 * A few mei fixes; * Some improvements in D3; * Support for new FW API commands; * Fixes and cleanups in device configurations; * Support some new FW API command versions; * Fix WGDS revision 3 reading bug; * Some firmware debugging improvements; * Fixes for in device configuration structures; * Improvements in the session protection code; * Support SAR GEO Offset Mapping (SGOM) via BIOS; * Continued work on the new Bz device family; * Some more firmware debugging improvements; * Support new FW API version 68; * Add some new device IDs; * Some other small fixes, clean-ups and improvements.
-
- 15 Dec, 2021 1 commit
-
-
zhangyue authored
Limit the max of 'ii'. If 'ii' greater than or equal to 'RSI_MAX_VIFS', the array 'adapter->vifs' may be out of bound Signed-off-by:
zhangyue <zhangyue1@kylinos.cn> Signed-off-by:
-
- 14 Dec, 2021 15 commits
-
-
David Mosberger-Tang authored
This follows normal Linux convention and is more useful since the new name will make it apparent which network device the work-queue is for (e.g., the name will be "wlan0-wq" for network device "wlan0"). hif_workqueue allocation has to move from cfg80211.c:wilc_cfg80211_init() to netdev.c:wilc_netdev_ifc_init() because the network device name is not known until after the netdev is registered. The move also makes sense because netdev.c is already responsible for destroying the work queue when it is no longer needed. Signed-off-by:
-
David Mosberger-Tang authored
This follows normal Linux conventions and is also more useful because the netdevice name is part of the task name (e.g., "wlan0-tx" for network device "wlan0"). Signed-off-by:
-
David Mosberger-Tang authored
This change follows normal Linux convention and the new name is more useful since it'll be clear which irq handler statistics correspond to which net device. Signed-off-by:
-
David Mosberger-Tang authored
The name "wilc1000_spi" follows normal Linux conventions and also is analogous to the SDIO driver, which uses "wilc1000_sdio". Signed-off-by:
-
David Mosberger-Tang authored
When deinitializing the driver, one or more "FW not responding" error appears on the console. This appears to be due to wilc_wlan_stop() disabling host/WILC1000 communication, but then right afterwards, it tries to release the bus with chip-sleep enabled. The problem is enabling the chip-sleep cannot success once host/WILC1000 communication is disabled. Fix by only releasing the bus. Signed-off-by:
-
David Mosberger-Tang authored
The USE_SPI_DMA macro name suggests that it could be set to 1 to control whether or not SPI DMA should be used. However, that's not what it does. If set to 1, it'll set the SPI messages' "is_dma_mapped" flag to true, even though the tx/rx buffers aren't actually DMA mapped by the driver. In other words, setting this flag to 1 will break the driver. Best to clean up this confusion by removing the macro altogether. There is no need to explicitly initialize "is_dma_mapped" because the message is cleared to zero anyhow, so "is_dma_mapped" is set to false by default. Signed-off-by:
Ajay Singh <ajay.kathat@microchip.com> Signed-off-by:
-
David Mosberger-Tang authored
Add missing newline in pr_err() message. Signed-off-by:
-
David Mosberger-Tang authored
The messages appears to have been copied from wilc_get_mac_address and says "get" when it should say "set". Signed-off-by:
-
Ching-Te Ku authored
Update COEX version. Signed-off-by:
Ching-Te Ku <ku920601@realtek.com> Signed-off-by:
Ping-Ke Shih <pkshih@realtek.com> Signed-off-by:
-
Ching-Te Ku authored
It's unnecessary to leave WL PS while C2H comes. Signed-off-by:
-
Ching-Te Ku authored
Move _chk_btc_err and update_bt_cnt to _chk_btc_report(), so we can update counter/info to COEX at a proper moment, instead of relying on a user does cat the debug info periodically. Signed-off-by:
-
Ching-Te Ku authored
To distinguish three types of LPS state. Signed-off-by:
-
Ching-Te Ku authored
Add function to get and parse BT polluted counter. When WLAN Tx was dropped by BT, the packet will be marked as BT polluted. Signed-off-by:
-
Ching-Te Ku authored
Prevent to send H2C request to FW when BTC is not initialized or WL is under power saving. Add counter to count the H2C success or fail. Signed-off-by:
-
Ching-Te Ku authored
To resolve C2H handle length mismatch, or it will parse the c2h content out of array. Signed-off-by:
-
- 13 Dec, 2021 1 commit
-
-
Loic Poulain authored
cpu_to_le16 has been used for all keep-alive fields, but 'perdio_msec' is a 32-bit field and 'keep_alive_id' a 8-bit one. Fix that. Fixes: 7a6cfe28 ("brcmfmac: Configure keep-alive packet on suspend") Reported-by:
kernel test robot <lkp@intel.com> Signed-off-by:
Loic Poulain <loic.poulain@linaro.org> Signed-off-by:
-
- 08 Dec, 2021 20 commits
-
-
Kees Cook authored
The earlier __packed annotations added in commit d71038c0 ("libertas: Fix alignment issues in libertas core") were not duplicated when libertas_af was added in commit 7670e62c ("libertas_tf: header file"), even though they share several structure definitions. Add the missing annotations which commit 642a5747 ("libertas_tf: Use struct_group() for memcpy() region") exposed. Quoting the prior libertas fix: "Data structures that come over the wire from the WLAN firmware must be packed." Reported-by:
Kees Cook <keescook@chromium.org> Signed-off-by:
-
Kees Cook authored
Build testing of the newly added struct_group() usage missed smaller architecture width tests for changes to pahole output. Add the missed __packed annotation with struct_group() usage in a __packed struct. Reported-by:
-
Brian Norris authored
Quoting Jia-Ju Bai <baijiaju1990@gmail.com>: mwifiex_dequeue_tx_packet() spin_lock_bh(&priv->wmm.ra_list_spinlock); --> Line 1432 (Lock A) mwifiex_send_addba() spin_lock_bh(&priv->sta_list_spinlock); --> Line 608 (Lock B) mwifiex_process_sta_tx_pause() spin_lock_bh(&priv->sta_list_spinlock); --> Line 398 (Lock B) mwifiex_update_ralist_tx_pause() spin_lock_bh(&priv->wmm.ra_list_spinlock); --> Line 941 (Lock A) Similar report for mwifiex_process_uap_tx_pause(). While the locking expectations in this driver are a bit unclear, the Fixed commit only intended to protect the sta_ptr, so we can drop the lock as soon as we're done with it. IIUC, this deadlock cannot actually happen, because command event processing (which calls mwifiex_process_sta_tx_pause()) is sequentialized with TX packet processing (e.g., mwifiex_dequeue_tx_packet()) via the main loop (mwifiex_main_process()). But it's good not to leave this potential issue lurking. Fixes: f0f7c227 ("mwifiex: minor cleanups w/ sta_list_spinlock in cfg80211.c") Cc: Douglas Anderson <dianders@chromium.org> Reported-by:TOTE Robot <oslab@tsinghua.edu.cn> Link: https://lore.kernel.org/linux-wireless/0e495b14-efbb-e0da-37bd-af6bd677ee2c@gmail.com/Signed-off-by:
Brian Norris <briannorris@chromium.org> Reviewed-by:
Douglas Anderson <dianders@chromium.org> Signed-off-by:
-
David Mosberger-Tang authored
This eliminates warning message: SPI driver WILC_SPI has no spi_device_id for microchip,wilc1000 and makes device-tree autoloading work. Signed-off-by:
-
Ping-Ke Shih authored
This is found by Smatch static checker warning: drivers/net/wireless/realtek/rtw89/mac80211.c:31 rtw89_ops_tx() error: uninitialized symbol 'qsel'. The warning is because 'qsel' isn't filled by rtw89_core_tx_write() due to failed to write. The way to fix it is to avoid kicking off TX DMA, so add 'return' to the failure case. Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Zong-Zhe Yang authored
Originally, cch_by_bw recorded center channels of each available bandwidths under current bandwidth. And the plan was to iterate cch_by_bw as parameters to query other configurations. However, we have not used it for the time being. Keeping it will disturb the follow-up things, such as bandwidth 160 MHz, so we remove it for now. If it's really needed at some point, we will redesign it. Signed-off-by:
Zong-Zhe Yang <kevin_yang@realtek.com> Signed-off-by:
-
Ping-Ke Shih authored
With wrong rtwsta->mac_id, it can't send out ack properly when we receive assoc response occasionally. Then, it failed to connect an AP. The cause is that we store 'sta' and use it somewhere. To correct this, remove the variable and use mac_id in drv_priv of 'sta' or 'vif' passed by mac80211. Signed-off-by:
-
Zong-Zhe Yang authored
Show channel and bandwidth in debugfs tx_pwr_tbl to make it easier to understand which tx power table is shown currently, and to reduce additional checks through other ways. Signed-off-by:
-
Yan-Hsuan Chuang authored
It is useful to fix the bit rate of TX packets. For example, if someone is measuring the TX power, or debugging with the issues of the TX throughput on the field. To set the value of fixed rate, one should input corresponding desc rate index (ex, 0x0b for DESC_RATE54M to fix at 54 Mbps). Set a value larger than DESC_RATE_MAX will disable fix rate, so the rate adaptive mechanism can resume to work. Example, To fix rate at MCS 1: echo 0x0d > /sys/kernel/debug/ieee80211/phy0/rtw88/fix_rate To not to fix rate: echo 0xff > /sys/kernel/debug/ieee80211/phy0/rtw88/fix_rate To know which rate was fixed at: cat /sys/kernel/debug/ieee80211/phy0/rtw88/fix_rate Signed-off-by:
Yan-Hsuan Chuang <yhchuang@realtek.com> Signed-off-by:
-
Po Hao Huang authored
We used to fill in rx skbs' frequency field by mac80211's current channel value. In some cases, mac80211 switches channel before all rx packets have been processed. This results in incorrect bss info. We fix this by filling in frequency field with channel index obtained from hardware, then fix potential cck missing issue by skb's original hw rate. After all fix is done, convert hw rate back to the supported band rate index. Signed-off-by:
Po Hao Huang <phhuang@realtek.com> Signed-off-by:
-
Ping-Ke Shih authored
Update scan_mac_addr to address CAM as A1, so hardware can ACK probe response properly. Signed-off-by:
-
Ping-Ke Shih authored
In order to use compiler to check if we do improper cast of const* on arguments of inline function. Signed-off-by:
Takashi Iwai <tiwai@suse.de> Tested-by:
Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by:
-
Ping-Ke Shih authored
Add 'const' to be clear that this is a read-only access, and this patch doesn't change logic at all. Signed-off-by:
-
Jakub Kicinski authored
Alexandra Winter says: ==================== s390/net: updates 2021-12-06 This brings some maintenance improvements and removes some unnecessary code checks. ==================== Link: https://lore.kernel.org/r/20211207090452.1155688-1-wintera@linux.ibm.comSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Julian Wiedmann authored
If qeth_check_outbound_queue() finds a partially filled TX buffer on the queue and flushes it, then the queue _must_ have been in packing mode. Remove the redundant check when updating the relevant statistics. Signed-off-by:
Julian Wiedmann <jwi@linux.ibm.com> Reviewed-by:
Alexandra Winter <wintera@linux.ibm.com> Signed-off-by:
-
Julian Wiedmann authored
Avoid a conditional branch for L2 devices when selecting the TX queue, and have shared logic for OSA devices. Signed-off-by:
-
Julian Wiedmann authored
qeth_l2_detect_dev2br_support() will only set brport_hw_features for IQD devices. So qeth_l2_bridge_getlink() and qeth_l2_bridge_setlink() will always return -EOPNOTSUPP on OSA devices. Just don't offer these callbacks instead. Signed-off-by:
-
Julian Wiedmann authored
Splitting up the netdev_ops allows for fine-tuning some of the ndo's in subsequent patches. Signed-off-by:
Wenjia Zhang <wenjia@linux.ibm.com> Signed-off-by:
-
Julian Wiedmann authored
Now that the OSN code is gone, we don't need the second switch statement in the RX path. And getting rid of its (unreachable) default case is a nice simplification. Also don't pass in the full HW header, all we still need is a flag to indicate whether the skb can use CSO. This we can already obtain during the first peek at the HW header. Signed-off-by:
-
Kees Cook authored
When building under -Warray-bounds, the compiler is especially conservative when faced with casts from a smaller object to a larger object. While this has found many real bugs, there are some cases that are currently false positives (like here). With this as one of the last few instances of the warning in the kernel before -Warray-bounds can be enabled globally, rearrange the functions so that there is a header-only version of hvs_send_data(). Silences this warning: net/vmw_vsock/hyperv_transport.c: In function 'hvs_shutdown_lock_held.constprop': net/vmw_vsock/hyperv_transport.c:231:32: warning: array subscript 'struct hvs_send_buf[0]' is partly outside array bounds of 'struct vmpipe_proto_header[1]' [-Warray-bounds] 231 | send_buf->hdr.pkt_type = 1; | ~~~~~~~~~~~~~~~~~~~~~~~^~~ net/vmw_vsock/hyperv_transport.c:465:36: note: while referencing 'hdr' 465 | struct vmpipe_proto_header hdr; | ^~~ This change results in no executable instruction differences. Signed-off-by:Wei Liu <wei.liu@kernel.org> Link: https://lore.kernel.org/r/20211207063217.2591451-1-keescook@chromium.orgSigned-off-by:
-
