- 05 Dec, 2023 8 commits
-
-
Kang Yang authored
For AP mode, the peer is created earlier in ath12k_mac_op_add_interface() but ath12k_mac_op_assign_vif_chanctx() will try to create peer again. Then an error will return which makes AP mode startup fail. Kernel log: [ 5017.665006] ath12k_pci 0000:04:00.0: failed to create peer after vdev start delay: -22 wpa_supplicant log: Failed to set beacon parameters Interface initialization failed wls1: interface state UNINITIALIZED->DISABLED wls1: AP-DISABLED wls1: Unable to setup interface. Failed to initialize AP interface wls1: interface state DISABLED->DISABLED wls1: AP-DISABLED So fix this check and enable AP mode for WCN7850, as now AP mode works normally. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Signed-off-by:
Kang Yang <quic_kangyang@quicinc.com> Acked-by:
Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by:
Kalle Valo <quic_kvalo@quicinc.com> Link: https://lore.kernel.org/r/20231121022459.17209-1-quic_kangyang@quicinc.com
-
Kang Yang authored
With VT-d disabled on Intel platform, ath12k gets only one MSI vector. In that case, ath12k does not free IRQ when doing suspend, hence the kernel has to migrate it to CPU0 (if it was affine to other CPUs) and allocates a new MSI vector. However, ath12k has no chance to reconfig it to HW srngs during this phase, thus ath12k fails to resume. This issue can be fixed by setting IRQ affinity to CPU0 before request_irq is called. With such affinity, migration will not happen and thus the vector keeps unchanged during suspend/resume. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Signed-off-by:
-
Kang Yang authored
Current code enables ASPM by default, it allows MHI to enter M2 state. In case of one MSI vector, system hang is observed if ath12k does MHI register reading in this state. The workaround here is to prevent MHI from entering M2 state, this can be done by disabling ASPM if only one MSI vector is used. When using 32 vectors ASPM is enabled as before. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Signed-off-by:
-
Kang Yang authored
On some platforms it's not possible to allocate 32 MSI vectors for various reasons, maybe kernel configuration, VT-d disabled, buggy BIOS etc. So ath12k was not able to use WCN7850 PCI devices on those platforms. Add support for one MSI vector to solve that. In case of one MSI vector, interrupt migration needs to be disabled. This is because when interrupt migration happens, the msi_data may change. However, msi_data is already programmed to rings during initial phase and ath12k has no way to know that msi_data is changed during run time and reprogram again. In case of one MSI vector, MHI subsystem should not use IRQF_NO_SUSPEND as WCN7850 doesn't set this flag too. Ath12k doesn't need to leave IRQ enabled in suspend state. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Signed-off-by:
-
Kang Yang authored
This is to prepare for one MSI vector support. IRQ enable and disable of CE and DP are done only in case of multiple MSI vectors. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Signed-off-by:
-
Kang Yang authored
Like ATH12K_PCI_IRQ_CE0_OFFSET, define ATH12K_PCI_IRQ_DP_OFFSET for DP to save the IRQ instead of base_vector from MSI config. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Signed-off-by:
-
Kang Yang authored
Add two flags to indicate whether IRQ handler for CE and DP can be called. This is because in one MSI vector case, interrupt is not disabled in hif_stop and hif_irq_disable. So if interrupt is disabled, MHI interrupt is disabled too. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Signed-off-by:
-
Kang Yang authored
The reservation mode of interrupts in kernel assigns a dummy vector when the interrupt is allocated and assigns a real vector when the request_irq is called. The reservation mode helps to ease vector pressure when devices with a large amount of queues/interrupts are initialized, but only a minimal subset of those queues/interrupts is actually used. So on reservation mode, the msi_data may change after request_irq is called, then it will lead to spurious interrupt. But when VT-d in BIOS is enabled and ath12k can get 32 MSI vectors, ath12k always get the same msi_data before and after request_irq. So in case of one MSI vector, ath12k need read msi_data again after request_irq is called, and then the correct msi_data is programmed into WCN7850 hardware components. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4 Signed-off-by:
-
- 01 Dec, 2023 5 commits
-
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Reviewed-by:
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Karthikeyan Periyasamy authored
Currently, the helper functions are accessing mac80211 hw data from the radio (ar) structure repeatedly. So optimize these helper functions by storing mac80211 hw data locally and accessing it directly. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.0.1-00029-QCAHKSWPL_SILICONZ-1 Signed-off-by:
Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Acked-by:
-
Karthikeyan Periyasamy authored
Currently mac80211 hw data is accessed by convert the hw to radio (ar) structure and then radio to hw structure which is not necessary in some places where mac80211 hw data is already present. So in that kind of places avoid the conversion and directly access the mac80211 hw data. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.0.1-00029-QCAHKSWPL_SILICONZ-1 Signed-off-by:
-
Muna Sinada authored
Add 320 MHz bandwidth as a new bandwidth enum for ATH12k driver. This is extending existing bandwidth related enums to include 320 MHz. This is a precursor to supporting 320 MHz in the future. Sanity test performed to confirm that there was no impact in existing bandwidths. Additionally update QuIC copyright to include 2023 in hal_rx.h. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00125-QCAHKSWPL_SILICONZ-1 Signed-off-by:
Muna Sinada <quic_msinada@quicinc.com> Acked-by:
-
- 30 Nov, 2023 8 commits
-
-
Uwe Kleine-König authored
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove(). Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Signed-off-by:
-
Baochen Qiang authored
We are seeing below error randomly in the case where only one MSI vector is configured: kernel: ath11k_pci 0000:03:00.0: wmi command 16387 timeout The reason is, currently, in ath11k_pcic_ext_irq_enable(), ATH11K_FLAG_EXT_IRQ_ENABLED is set before NAPI is enabled. This results in a race condition: after ATH11K_FLAG_EXT_IRQ_ENABLED is set but before NAPI enabled, CE interrupt breaks in. Since IRQ is shared by CE and data path, ath11k_pcic_ext_interrupt_handler() is also called where we call disable_irq_nosync() to disable IRQ. Then napi_schedule() is called but it does nothing because NAPI is not enabled at that time, meaning ath11k_pcic_ext_grp_napi_poll() will never run, so we have no chance to call enable_irq() to enable IRQ back. Finally we get above error. Fix it by setting ATH11K_FLAG_EXT_IRQ_ENABLED after all NAPI and IRQ work are done. With the fix, we are sure that by the time ATH11K_FLAG_EXT_IRQ_ENABLED is set, NAPI is enabled. Note that the fix above also introduce some side effects: if ath11k_pcic_ext_interrupt_handler() breaks in after NAPI enabled but before ATH11K_FLAG_EXT_IRQ_ENABLED set, nothing will be done by the handler this time, the work will be postponed till the next time the IRQ fires. Tested-on: WCN6855 hw2.1 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.23 Signed-off-by:
Baochen Qiang <quic_bqiang@quicinc.com> Acked-by:
-
Jeff Johnson authored
The misspelled pauload member of struct ath11k_htc_record is unused, so remove it. Signed-off-by:
-
Jeff Johnson authored
Transform the zero-length arrays in ath10k_htc_record into proper flexible arrays via the DECLARE_FLEX_ARRAY() macro. This helps with ongoing efforts to globally enable -Warray-bounds. Signed-off-by:
Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by:
-
Jeff Johnson authored
The misspelled pauload member of struct ath10k_htc_record is unused, so remove it. Signed-off-by:
-
Jeff Johnson authored
Update the copyright for all ath10k files modified on behalf of Qualcomm Innovation Center, Inc. in 2021 through 2023. Signed-off-by:
-
Jeff Johnson authored
Update the copyright for all ath11k files modified on behalf of Qualcomm Innovation Center, Inc. in 2021 through 2023. Signed-off-by:
-
Jeff Johnson authored
Update the copyright for all ath12k files modified on behalf of Qualcomm Innovation Center, Inc. in 2023. Signed-off-by:
-
- 29 Nov, 2023 9 commits
-
-
Minsuk Kang authored
Fix an array-index-out-of-bounds read in ath9k_htc_txstatus(). The bug occurs when txs->cnt, data from a URB provided by a USB device, is bigger than the size of the array txs->txstatus, which is HTC_MAX_TX_STATUS. WARN_ON() already checks it, but there is no bug handling code after the check. Make the function return if that is the case. Found by a modified version of syzkaller. UBSAN: array-index-out-of-bounds in htc_drv_txrx.c index 13 is out of range for type '__wmi_event_txstatus [12]' Call Trace: ath9k_htc_txstatus ath9k_wmi_event_tasklet tasklet_action_common __do_softirq irq_exit_rxu sysvec_apic_timer_interrupt Signed-off-by:
Minsuk Kang <linuxlovemin@yonsei.ac.kr> Acked-by:
Toke Høiland-Jørgensen <toke@toke.dk> Signed-off-by:
-
Wu Yunchuan authored
No need cast (void *) to (struct owl_ctx *), (struct ath_hw *), (struct cmd_buf *) or other types. Signed-off-by:
Wu Yunchuan <yunchuan@nfschina.com> Reviewed-by:
-
Karthikeyan Periyasamy authored
Currently data path Rxdma ring structure store the IDR buffer and lock. These IDR handling is needed only for SW cookie conversion and not needed for HW cookie conversion. REO Rxdma ring use the HW cookie conversion and monitor Rxdma ring use the SW cookie conversion. Since idr not needed for REO Rxdma ring, remove the IDR data entity from the data path Rxdma ring structure. Introduce the new data path ring structure for monitor rxmda rings since it need IDR data entity. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00125-QCAHKSWPL_SILICONZ-1 Signed-off-by:
-
Karthikeyan Periyasamy authored
Currently Rxdma replenish require HW conversion argument which is unnecessary argument since ath12k driver configures the Rxdma only in HW conversion. To optimize the rx data path per packet, avoid the explicit unnecessary argument and condition check in the rx replenish. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00125-QCAHKSWPL_SILICONZ-1 Signed-off-by:
-
Karthikeyan Periyasamy authored
Currently all Rxdma replenish callers pass the same return buffer manager id argument, so make it implicitly. To optimize the rx data path per packet, avoid the explicit unnecessary argument in Rxdma replenish function. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00125-QCAHKSWPL_SILICONZ-1 Signed-off-by:
-
Karthikeyan Periyasamy authored
Currently all Rxdma replenish callers pass zero for the mac id argument, so make it as zero implicitly. To optimize the rx data path per packet, avoid the explicit unnecessary argument in Rxdma replenish function. Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00125-QCAHKSWPL_SILICONZ-1 Signed-off-by:
-
Karthikeyan Periyasamy authored
When the core rfkill config throws error, it should free the allocated resources. Currently it is not freeing the core pdev create resources. Avoid this issue by calling the core pdev destroy in the error handler of core rfkill config. Found this issue in the code review and it is compile tested only. Fixes: 004ccbc0 ("wifi: ath12k: add support for hardware rfkill for WCN7850") Signed-off-by:
-
James Prestwood authored
Let ath12k select this option automatically which makes building more intuitive if the user enables this driver (rather than the driver not building unless CRYPTO_MICHAEL_MIC is explicitly enabled). Further investigation shows that ath11k and ath12k are the only who use 'depends on' with CRYPTO_MICHAEL_MIC: ./drivers/net/wireless/intel/ipw2x00/Kconfig: select CRYPTO_MICHAEL_MIC ./drivers/net/wireless/intersil/hostap/Kconfig: select CRYPTO_MICHAEL_MIC ./drivers/net/wireless/intersil/orinoco/Kconfig: select CRYPTO_MICHAEL_MIC ./drivers/net/wireless/ath/ath11k/Kconfig: depends on CRYPTO_MICHAEL_MIC ./drivers/net/wireless/ath/ath12k/Kconfig: depends on CRYPTO_MICHAEL_MIC ./drivers/staging/rtl8192e/Kconfig: select CRYPTO_MICHAEL_MIC ./drivers/staging/ks7010/Kconfig: select CRYPTO_MICHAEL_MIC Signed-off-by:
James Prestwood <prestwoj@gmail.com> Signed-off-by:
-
James Prestwood authored
Let ath11k select this option automatically which makes building more intuitive if the user enables this driver (rather than the driver not building unless CRYPTO_MICHAEL_MIC is explicitly enabled). Further investigation shows that ath11k and ath12k are the only who use 'depends on' with CRYPTO_MICHAEL_MIC: ./drivers/net/wireless/intel/ipw2x00/Kconfig: select CRYPTO_MICHAEL_MIC ./drivers/net/wireless/intersil/hostap/Kconfig: select CRYPTO_MICHAEL_MIC ./drivers/net/wireless/intersil/orinoco/Kconfig: select CRYPTO_MICHAEL_MIC ./drivers/net/wireless/ath/ath11k/Kconfig: depends on CRYPTO_MICHAEL_MIC ./drivers/net/wireless/ath/ath12k/Kconfig: depends on CRYPTO_MICHAEL_MIC ./drivers/staging/rtl8192e/Kconfig: select CRYPTO_MICHAEL_MIC ./drivers/staging/ks7010/Kconfig: select CRYPTO_MICHAEL_MIC Signed-off-by:
-
- 22 Nov, 2023 8 commits
-
-
Jeff Johnson authored
Currently wmi.h has two separate set of definitions for peer flags. One set of flags is defined in enum wmi_tlv_peer_flags, and, except for the last three, are named WMI_TLV_PEER_*. The other set of flags are defined as macros, and are named WMI_PEER_*. The last three macros have the same name as the last three wmi_tlv_peer_flags enumerators. The code only uses the WMI_PEER_* names; the WMI_TLV_PEER_* names are unused. So as a first step in consolidation, remove all the WMI_TLV_PEER_* names. But since having an enum to define all the flags is actually a good thing since that provides a handle by which to refer to the entire set of flags, recast the WMI_PEER_* macros into enumerators. Compile tested only. Signed-off-by:
-
Jeff Johnson authored
Currently wmi.h has two separate set of definitions for peer flags. One set of flags is defined in enum wmi_tlv_peer_flags, and, except for the last three, are named WMI_TLV_PEER_*. The other set of flags are defined as macros, and are named WMI_PEER_*. The last three macros have the same name as the last three wmi_tlv_peer_flags enumerators. The code only uses the WMI_PEER_* names; the WMI_TLV_PEER_* names are unused. So as a first step in consolidation, remove all the WMI_TLV_PEER_* names. But since having an enum to define all the flags is actually a good thing since that provides a handle by which to refer to the entire set of flags, recast the WMI_PEER_* macros into enumerators. Compile tested only. Signed-off-by:
-
Jeff Johnson authored
Currently both struct ath12k_wmi_pdev and struct ath12k_wmi_base define: const struct wmi_peer_flags_map *peer_flags; But that member is not used, and in fact, struct wmi_peer_flags_map is not defined within ath12k; these are obsolete remnants inherited from ath11k. So remove them. Compile tested only. Signed-off-by: -
Jeff Johnson authored
Currently both struct ath11k_pdev_wmi and struct ath11k_wmi_base define: const struct wmi_peer_flags_map *peer_flags; But that member is not used, and in fact, struct wmi_peer_flags_map is not defined within ath11k; these are obsolete remnants inherited from ath10k. So remove them. Compile tested only. Signed-off-by:
-
Jeff Johnson authored
Currently struct ath12k defines the following member: struct ieee80211_ops *ops; This is being flagged by checkpatch.pl: WARNING: struct ieee80211_ops should normally be const The original plan was to add the const qualifier. However, it turns out this is actually unused, so remove it. No functional changes, compile tested only. Signed-off-by:
-
Jeff Johnson authored
Currently struct ath11k defines the following member: struct ieee80211_ops *ops; This is being flagged by checkpatch.pl: WARNING: struct ieee80211_ops should normally be const The original plan was to add the const qualifier. However, it turns out this is actually unused, so remove it. No functional changes, compile tested only. Signed-off-by: -
Jeff Johnson authored
struct ath10k_htc_frame is unused, and since it illogically contains two consecutive flexible arrays, it could never be used, so remove it. No functional changes, compile tested only. Signed-off-by:
-
Dmitry Antipov authored
Since 'ieee80211_txq_get_depth()' allows NULL for 2nd and 3rd arguments, simplify '__ath10k_htt_tx_txq_recalc()' by dropping unused 'frame_cnt'. Compile tested only. Signed-off-by:
Dmitry Antipov <dmantipov@yandex.ru> Acked-by:
-
- 13 Nov, 2023 2 commits
-
-
Yang Li authored
./drivers/net/wireless/ath/ath11k/fw.c:136:2-3: Unneeded semicolon Reported-by:
Abaci Robot <abaci@linux.alibaba.com> Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=7073Signed-off-by:
Yang Li <yang.lee@linux.alibaba.com> Acked-by:
-
Luca Weiss authored
If we already have gotten the rproc_handle (meaning the "qcom,rproc" property is defined in the devicetree), it's a valid state that the remoteproc module hasn't probed yet so we should defer probing instead of just failing to probe. This resolves a race condition when the ath11k driver probes and fails before the wpss remoteproc driver has probed, like the following: [ 6.232360] ath11k 17a10040.wifi: failed to get rproc [ 6.232366] ath11k 17a10040.wifi: failed to get rproc: -22 [ 6.232478] ath11k: probe of 17a10040.wifi failed with error -22 ... [ 6.252415] remoteproc remoteproc2: 8a00000.remoteproc is available [ 6.252776] remoteproc remoteproc2: powering up 8a00000.remoteproc [ 6.252781] remoteproc remoteproc2: Booting fw image qcom/qcm6490/fairphone5/wpss.mdt, size 7188 So, defer the probe if we hit that so we can retry later once the wpss remoteproc is available. Tested-on: WCN6750 hw1.0 AHB WLAN.MSL.1.0.1-01264-QCAMSLSWPLZ-1.37886.3 Fixes: d5c65159 ("ath11k: driver for Qualcomm IEEE 802.11ax devices") Signed-off-by:Luca Weiss <luca.weiss@fairphone.com> Signed-off-by:
-
