The DMA buffers of dbring which is used for spectral/cfr
starts at certain offset from original kmalloc() returned buffer.
This is not cache line aligned.
And also driver tries to access the data that is immediately before
this offset address (i.e. buff->paddr) after doing dma map.
This will cause cache line sharing issues and data corruption,
if CPU happen to write back cache after HW has dma'ed the data.
Fix this by mapping a cache line aligned buffer to dma.
Tested on: IPQ8074 hw2.0 AHB WLAN.HK.2.5.0.1-01100-QCAHKSWPL_SILICONZ-1
Signed-off-by: Rameshkumar Sundaram <quic_ramess@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1635831693-15962-1-git-send-email-quic_ramess@quicinc.com
Firmware credit flow control is enabled for WMI control services,
which expects available tokens should be acquired before sending a
command to the target. Also the token gets released when firmware
receives the command.
This credit-based flow limits driver to send WMI command only
when the token available which is causing WMI commands to timeout and
return -EAGAIN, whereas firmware has enough capability to process the
WMI command. To fix this Tx starvation issue, introduce the ability to
disable the credit flow for the WMI path.
The driver sends WMI configuration for disabling credit flow to firmware
by two ways.
1. By using a global flag
(HTC_MSG_SETUP_COMPLETE_EX_ID msg type flags)
2. By using a local flag
(ATH11K_HTC_CONN_FLAGS_DISABLE_CREDIT_FLOW_CTRL = 1 << 3)
Ath11k uses both these configurations to disable credit flow for the
WMI path completely.
Also added a hw_param member for credit flow control by which we can
enable or disable it based on per-target basis. Currently we are disabling
credit flow for IPQ8074, IPQ6018, and QCN9074 as recommended by firmware.
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.4.0.1-01492-QCAHKSWPL_SILICONZ-1
Tested-on: IPQ6018 hw1.0 AHB WLAN.HK.2.4.0.1-00330-QCAHKSWPL_SILICONZ-1
Co-developed-by: Pravas Kumar Panda <kumarpan@codeaurora.org>
Signed-off-by: Pravas Kumar Panda <kumarpan@codeaurora.org>
Signed-off-by: P Praneesh <quic_ppranees@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1635156494-20059-1-git-send-email-quic_ppranees@quicinc.com
Some ETSI countries have a small overlap in the wireless-regdb with an ETSI
channel (5590-5650). A good example is Australia:
country AU: DFS-ETSI
(2400 - 2483.5 @ 40), (36)
(5150 - 5250 @ 80), (23), NO-OUTDOOR, AUTO-BW
(5250 - 5350 @ 80), (20), NO-OUTDOOR, AUTO-BW, DFS
(5470 - 5600 @ 80), (27), DFS
(5650 - 5730 @ 80), (27), DFS
(5730 - 5850 @ 80), (36)
(57000 - 66000 @ 2160), (43), NO-OUTDOOR
If the firmware (or the BDF) is shipped with these rules then there is only
a 10 MHz overlap with the weather radar:
* below: 5470 - 5590
* weather radar: 5590 - 5600
* above: (none for the rule "5470 - 5600 @ 80")
There are several wrong assumption in the ath11k code:
* there is always a valid range below the weather radar
(actually: there could be no range below the weather radar range OR range
could be smaller than 20 MHz)
* intersected range in the weather radar range is valid
(actually: the range could be smaller than 20 MHz)
* range above weather radar is either empty or valid
(actually: the range could be smaller than 20 MHz)
These wrong assumption will lead in this example to a rule
(5590 - 5600 @ 20), (N/A, 27), (600000 ms), DFS, AUTO-BW
which is invalid according to is_valid_reg_rule() because the freq_diff is
only 10 MHz but the max_bandwidth is set to 20 MHz. Which results in a
rejection like:
WARNING: at backports-20210222_001-4.4.60-b157d2276/net/wireless/reg.c:3984
[...]
Call trace:
[<ffffffbffc3d2e50>] reg_get_max_bandwidth+0x300/0x3a8 [cfg80211]
[<ffffffbffc3d3d0c>] regulatory_set_wiphy_regd_sync+0x3c/0x98 [cfg80211]
[<ffffffbffc651598>] ath11k_regd_update+0x1a8/0x210 [ath11k]
[<ffffffbffc652108>] ath11k_regd_update_work+0x18/0x20 [ath11k]
[<ffffffc0000a93e0>] process_one_work+0x1f8/0x340
[<ffffffc0000a9784>] worker_thread+0x25c/0x448
[<ffffffc0000aedc8>] kthread+0xd0/0xd8
[<ffffffc000085550>] ret_from_fork+0x10/0x40
ath11k c000000.wifi: failed to perform regd update : -22
Invalid regulatory domain detected
To avoid this, the algorithm has to be changed slightly. Instead of
splitting a rule which overlaps with the weather radar range into 3 pieces
and accepting the first two parts blindly, it must actually be checked for
each piece whether it is a valid range. And only if it is valid, add it to
the output array.
When these checks are in place, the processed rules for AU would end up as
country AU: DFS-ETSI
(2400 - 2483 @ 40), (N/A, 36), (N/A)
(5150 - 5250 @ 80), (6, 23), (N/A), NO-OUTDOOR, AUTO-BW
(5250 - 5350 @ 80), (6, 20), (0 ms), NO-OUTDOOR, DFS, AUTO-BW
(5470 - 5590 @ 80), (6, 27), (0 ms), DFS, AUTO-BW
(5650 - 5730 @ 80), (6, 27), (0 ms), DFS, AUTO-BW
(5730 - 5850 @ 80), (6, 36), (N/A), AUTO-BW
and will be accepted by the wireless regulatory code.
Fixes: d5c65159f2 ("ath11k: driver for Qualcomm IEEE 802.11ax devices")
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211112153116.1214421-1-sven@narfation.org
There is a race condition whereby the tx completion handler can be invoked
before the 'num_pending_mgmt_tx" count is incremented. If that occurs, we
could get warning trace indicating that 'num_pending_mgmt_tx' is 0 (because
it was not yet incremented). Ideally, this trace should be seen only if
mgmt tx has not happened but tx completion is received, and it is not
expected in this race condition.
Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.4.0.1-01386-QCAHKSWPL_SILICONZ-1
Co-developed-by: Lavanya Suresh <lavaks@codeaurora.org>
Signed-off-by: Lavanya Suresh <lavaks@codeaurora.org>
Signed-off-by: P Praneesh <quic_ppranees@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1635168282-8845-1-git-send-email-quic_ppranees@quicinc.com
In radio tap header, BAD FCS flag is not updated properly because
driver failed to update FCS_ERR flag in monitor mode.
In rx_desc, FCS_ERR information is available in rx_attention
structure and presence of this field indicates corresponding frame
failed FCS check.
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.4.0.1-01695-QCAHKSWPL_SILICONZ-1
Signed-off-by: P Praneesh <quic_ppranees@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1635164060-18423-1-git-send-email-quic_ppranees@quicinc.com
When station connect to AP, the wmi command sequence is:
peer_create->vdev_start->vdev_up
and sequence of station disconnect fo AP is:
peer_delete->vdev_down->vdev_stop
The sequence of disconnect is not opposite of connect, it caused firmware
crash when it handle wmi vdev stop cmd when the AP is support TWT of
802.11 ax, because firmware need access the bss peer for vdev stop cmd.
[ 390.438564] ath11k_pci 0000:05:00.0: wmi cmd send 0x6001 ret 0
[ 390.438567] ath11k_pci 0000:05:00.0: WMI peer create vdev_id 0 peer_addr c4:04:15:3b:e0:39
[ 390.472724] ath11k_pci 0000:05:00.0: mac vdev 0 start center_freq 2437 phymode 11ax-he20-2g
[ 390.472731] ath11k_pci 0000:05:00.0: wmi cmd send 0x5003 ret 0
[ 390.560849] ath11k_pci 0000:05:00.0: wmi cmd send 0x5005 ret 0
[ 390.560850] ath11k_pci 0000:05:00.0: WMI mgmt vdev up id 0x0 assoc id 1 bssid c4:04:15:3b:e0:39
[ 399.432896] ath11k_pci 0000:05:00.0: WMI peer delete vdev_id 0 peer_addr c4:04:15:3b:e0:39
[ 399.432902] ath11k_pci 0000:05:00.0: wmi cmd send 0x6002 ret 0
[ 399.441380] ath11k_pci 0000:05:00.0: wmi cmd send 0x5007 ret 0
[ 399.441381] ath11k_pci 0000:05:00.0: WMI vdev down id 0x0
[ 399.454681] ath11k_pci 0000:05:00.0: wmi cmd send 0x5006 ret 0
[ 399.454682] ath11k_pci 0000:05:00.0: WMI vdev stop id 0x0
The opposite sequence of disconnect should be:
vdev_down->vdev_stop->peer_delete
This patch change the sequence of disconnect for station as above
opposite sequence for QCA6390, firmware not crash again with this patch.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Wen Gong <quic_wgong@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211027093825.12167-1-quic_wgong@quicinc.com
Currently in ath11k_mac_op_sta_statistics() there is the following
logic:
if (!arsta->txrate.legacy && !arsta->txrate.nss)
return;
Unfortunately if this condition is true then the function returns without
setting parameters that follow the txrate. To address this issue remove the
return and instead invert the logic to set the txrate logic if
(arsta->txrate.legacy || arsta->txrate.nss).
The same was done also in ath10k in commit 1cd6ba8ae3 ("ath10k: remove return
for NL80211_STA_INFO_TX_BITRATE").
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211011084957.31024-1-wgong@codeaurora.org
When NSS ratio enabled reported by firmware, SUPPORTS_VHT_EXT_NSS_BW
is set in ath11k, meanwhile IEEE80211_VHT_EXT_NSS_BW_CAPABLE also
need to be set, otherwise it is invalid because spec in IEEE Std
802.11™‐2020 as below.
Table 9-273-Supported VHT-MCS and NSS Set subfields, it has subfield
VHT Extended NSS BW Capable, its definition is:
Indicates whether the STA is capable of interpreting the Extended NSS
BW Support subfield of the VHT Capabilities Information field.
dmesg have a message without this patch:
ieee80211 phy0: copying sband (band 1) due to VHT EXT NSS BW flag
It means mac80211 will set IEEE80211_VHT_EXT_NSS_BW_CAPABLE if ath11k not
set it in ieee80211_register_hw(). So it is better to set it in ath11k.
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-01720.1-QCAHSPSWPL_V1_V2_SILICONZ_LITE-1
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211013073704.15888-1-wgong@codeaurora.org
Avoid unnecessary idr_find calls before the idr_remove calls. Because
idr_remove gives the valid ptr if id is valid otherwise return NULL ptr.
So removed the idr_find before idr_remove in tx completion path. Also no
need to disable the bottom half preempt if it is already in the
bottom half context, so modify the spin_lock_bh to spin_lock in the
data tx completion path.
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.4.0.1-01734-QCAHKSWPL_SILICONZ-1 v2
Co-developed-by: Karthikeyan Periyasamy <periyasa@codeaurora.org>
Signed-off-by: Karthikeyan Periyasamy <periyasa@codeaurora.org>
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Signed-off-by: P Praneesh <ppranees@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1630560820-21905-13-git-send-email-ppranees@codeaurora.org
In datapath, add branch predictors where required in the process rx().
This protects high value rx path without having performance overhead.
Also while processing rx packets, the pointer that is returned by
rcu_dereference() is not dereferenced. so it is preferable to use
rcu_access_pointer() here.
Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.4.0.1.r2-00012-QCAHKSWPL_SILICONZ-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.4.0.1-01695-QCAHKSWPL_SILICONZ-1
Co-developed-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Signed-off-by: P Praneesh <ppranees@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1630560820-21905-8-git-send-email-ppranees@codeaurora.org
In ath11k_dp_process_rx(), after processing rx_desc from
ath11k_hal_srng_dst_get_next_entry(), ath11k_hal_srng_dst_num_free()
is accessed everytime because of done flag is not set.
To avoid this additional access to ath11k_hal_srng_dst_num_free(),
increment total_msdu_reaped only when continuation is not set and
update done flag correspondingly.
Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.4.0.1.r2-00012-QCAHKSWPL_SILICONZ-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.4.0.1-01695-QCAHKSWPL_SILICONZ-1
Co-developed-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Signed-off-by: P Praneesh <ppranees@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1630560820-21905-5-git-send-email-ppranees@codeaurora.org
tcl_data and reo_dst rings are currently being allocated using
dma_allocate_coherent() which is non cacheable.
Allocating ring memory from cacheable memory area allows cached descriptor
access and prefetch next descriptors to optimize CPU usage during
descriptor processing on NAPI. Based on the hardware param we can enable
or disable this feature for the corresponding platform.
Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.4.0.1.r2-00012-QCAHKSWPL_SILICONZ-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.4.0.1-01695-QCAHKSWPL_SILICONZ-1
Co-developed-by: Pradeep Kumar Chitrapu <pradeepc@codeaurora.org>
Signed-off-by: Pradeep Kumar Chitrapu <pradeepc@codeaurora.org>
Co-developed-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Signed-off-by: P Praneesh <ppranees@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1630560820-21905-3-git-send-email-ppranees@codeaurora.org
Currently the updated bandwidth for the peer will be configured
to the firmware after channel switch from the sta_rc_update_wk.
If the updated bandwidth is greater than the configured peer phymode
during the peer assoc may result firmware assert.
For example, initially AP is in HE40 mode and the peer phymode is
configured as MODE_11AX_HE40 during peer assoc. Now user change the
channel width to HE80 then, the peer bandwidth will be updated as
HE80 to the firmware.
This will trigger firmware assert due to peer bandwidth is greater than
the peer phymode.
Fix this issue by sending peer assoc command before setting the updated
peer bandwith to firmware.
Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.5.0.1-01100-QCAHKSWPL_SILICONZ-1
Signed-off-by: Venkateswara Naralasetty <quic_vnaralas@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1636644002-25446-1-git-send-email-quic_vnaralas@quicinc.com
In CE pipe cleanup, DMA memory gets freed by the aligned address
(base_addr_owner_space) which is wrong. It needs to be freed
by the address (base_addr_owner_space_unaligned) returned by
dma_alloc. So free the dma memory by the proper address.
This was found in code review.
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.5.0.1-00729-QCAHKSWPL_SILICONZ-3 v2
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1636561290-18966-1-git-send-email-quic_periyasa@quicinc.com
In STA kickout event processing, the peer object is protected
under spin lock BH. Release this lock after picking up the vdev_id
from the peer object instead after ieee80211_report_low_ack().
This will minimize the lock hold period which will improve
performance since base_lock is used across the data path.
This was found in code review.
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.5.0.1-00729-QCAHKSWPL_SILICONZ-3 v2
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1636560388-24955-1-git-send-email-quic_periyasa@quicinc.com
When there is an error in add interface process from
ath11k_mac_set_kickout(), the code attempts to handle a
fallback for add_interface. When this fallback succeeds, the
driver returns zero rather than error code. This leads to
success for the non created VAP. In cleanup, driver gets
remove interface callback for the non created VAP and
proceeds to self peer delete request which leads to FW assert.
Since it was already deleted on the fallback of add interface,
return the actual error code instead of fallback return code.
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.5.0.1-00729-QCAHKSWPL_SILICONZ-3 v2
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1636558557-2874-1-git-send-email-quic_periyasa@quicinc.com
Firmware expects host to authorize the peer after the successful key
install. But host authorize the peer before the key install, this trigger
the firmware assert which leads to Q6 crash. To avoid this Q6 crash, host
should authorize the peer after the key install. So introduce is_authorized
in peer object to identify that peer is authorize or not. When
IEEE80211_STA_CONTROL_PORT flag is unset, peer move to authorize state
before the vdev up. When the same flag is set then peer move to authorize
state after vdev up. So added authorise check in ath11k_bss_assoc() to
handle the earlier state transition case. Also added the WMI authorize
procedure in ath11k_mac_op_sta_state() to handle the non-earlier state
transition case.
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.4.0.1-01492-QCAHKSWPL_SILICONZ-1
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1636554200-12345-1-git-send-email-quic_periyasa@quicinc.com
AUTH flag is needed when peer assoc command is sent from host in
security mode for non-assoc cases. Firmware will handle AUTH flag
when client is associating as AUTH flag will be set after key exchange.
For internally provided peer assoc commands from host, there won't be
any key exchange, so AUTH flag is expected to be set in host.
Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.4.0.1-01838-QCAHKSWPL_SILICONZ-1
Co-developed-by: Lavanya Suresh <lavaks@codeaurora.org>
Signed-off-by: Lavanya Suresh <lavaks@codeaurora.org>
Signed-off-by: Rameshkumar Sundaram <quic_ramess@quicinc.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1635177786-20854-1-git-send-email-quic_ramess@quicinc.com
HTT_PPDU_STATS_CFG_PDEV_ID bit mask for target FW PPDU stats request message
was set as bit 8 to 15. Bit 8 is reserved for soc stats and pdev id starts from
bit 9. Hence change the bitmask as bit 9 to 15 and fill the proper pdev id in
the request message.
In commit 701e48a43e ("ath11k: add packet log support for QCA6390"), both
HTT_PPDU_STATS_CFG_PDEV_ID and pdev_mask were changed, but this pdev_mask
calculation is not valid for platforms which has multiple pdevs with 1 rxdma
per pdev, as this is writing same value(i.e. 2) for all pdevs. Hence fixed it
to consider pdev_idx as well, to make it compatible for both single and multi
pd cases.
Tested on: IPQ8074 hw2.0 AHB WLAN.HK.2.5.0.1-01092-QCAHKSWPL_SILICONZ-1
Tested on: IPQ6018 hw1.0 WLAN.HK.2.5.0.1-01067-QCAHKSWPL_SILICONZ-1
Fixes: 701e48a43e ("ath11k: add packet log support for QCA6390")
Co-developed-by: Sathishkumar Muruganandam <murugana@codeaurora.org>
Signed-off-by: Sathishkumar Muruganandam <murugana@codeaurora.org>
Signed-off-by: Rameshkumar Sundaram <ramess@codeaurora.org>
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210721212029.142388-10-jouni@codeaurora.org
Unexpected WDCMSG_TARGET_START replay can lead to null-ptr-deref
when ar->tx_cmd->odata is NULL. The patch adds a null check to
prevent such case.
KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
ar5523_cmd+0x46a/0x581 [ar5523]
ar5523_probe.cold+0x1b7/0x18da [ar5523]
? ar5523_cmd_rx_cb+0x7a0/0x7a0 [ar5523]
? __pm_runtime_set_status+0x54a/0x8f0
? _raw_spin_trylock_bh+0x120/0x120
? pm_runtime_barrier+0x220/0x220
? __pm_runtime_resume+0xb1/0xf0
usb_probe_interface+0x25b/0x710
really_probe+0x209/0x5d0
driver_probe_device+0xc6/0x1b0
device_driver_attach+0xe2/0x120
I found the bug using a custome USBFuzz port. It's a research work
to fuzz USB stack/drivers. I modified it to fuzz ath9k driver only,
providing hand-crafted usb descriptors to QEMU.
After fixing the code (fourth byte in usb packet) to WDCMSG_TARGET_START,
I got the null-ptr-deref bug. I believe the bug is triggerable whenever
cmd->odata is NULL. After patching, I tested with the same input and no
longer see the KASAN report.
This was NOT tested on a real device.
Signed-off-by: Zekun Shen <bruceshenzk@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/YXsmPQ3awHFLuAj2@10-18-43-117.dynapool.wireless.nyu.edu
To reduce power consumption enable 802.11 power save mode in station mode. This
allows both radio and CPU to sleep more.
Only enable the mode on QCA6390 and WCN6855, it's unknown how other hardware
families support this feature.
To test that power save mode is running run "iw dev wls1 set power_save off",
check there is no NULL Data frame seen by a sniffer. And run "iw dev wls1 set power_save
on" and check there is a NULL Data frame in sniffer.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211108123826.8463-2-kvalo@codeaurora.org
status.band is used in determination of status.rate -- for 5GHz on legacy
rates there is a linear shift between the BD descriptor's rate field and
the wcn36xx driver's rate table (wcn_5ghz_rates).
We have a special clause to populate status.band for hardware scan offload
frames. However, this block occurs after status.rate is already populated.
Correctly handle this dependency by moving the band block before the rate
block.
This patch addresses kernel warnings & missing scan results for 5GHz APs
that send their beacons/probe responses at the higher four legacy rates
(24-54 Mbps), when using hardware scan offload:
------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at net/mac80211/rx.c:4532 ieee80211_rx_napi+0x744/0x8d8
Modules linked in: wcn36xx [...]
CPU: 0 PID: 0 Comm: swapper/0 Tainted: G W 4.19.107-g73909fa #1
Hardware name: Square, Inc. T2 (all variants) (DT)
Call trace:
dump_backtrace+0x0/0x148
show_stack+0x14/0x1c
dump_stack+0xb8/0xf0
__warn+0x2ac/0x2d8
warn_slowpath_null+0x44/0x54
ieee80211_rx_napi+0x744/0x8d8
ieee80211_tasklet_handler+0xa4/0xe0
tasklet_action_common+0xe0/0x118
tasklet_action+0x20/0x28
__do_softirq+0x108/0x1ec
irq_exit+0xd4/0xd8
__handle_domain_irq+0x84/0xbc
gic_handle_irq+0x4c/0xb8
el1_irq+0xe8/0x190
lpm_cpuidle_enter+0x220/0x260
cpuidle_enter_state+0x114/0x1c0
cpuidle_enter+0x34/0x48
do_idle+0x150/0x268
cpu_startup_entry+0x20/0x24
rest_init+0xd4/0xe0
start_kernel+0x398/0x430
---[ end trace ae28cb759352b403 ]---
Fixes: 8a27ca3947 ("wcn36xx: Correct band/freq reporting on RX")
Signed-off-by: Benjamin Li <benl@squareup.com>
Tested-by: Loic Poulain <loic.poulain@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211104010548.1107405-2-benl@squareup.com
Right now we have a broken sequence where we enable DMA channel interrupts
which can be left enabled and never disabled if we hit an error path.
Worse still when we unload the driver, the DMA channel interrupt bits are
left intact. About the only saving grace here is that we do remember to
disable the wcnss interrupt when unload the driver.
Fixes: 8e84c25821 ("wcn36xx: mac80211 driver for Qualcomm WCN3660/WCN3680 hardware")
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211105122152.1580542-2-bryan.odonoghue@linaro.org
Pull networking updates from Jakub Kicinski:
"Core:
- Remove socket skb caches
- Add a SO_RESERVE_MEM socket op to forward allocate buffer space and
avoid memory accounting overhead on each message sent
- Introduce managed neighbor entries - added by control plane and
resolved by the kernel for use in acceleration paths (BPF / XDP
right now, HW offload users will benefit as well)
- Make neighbor eviction on link down controllable by userspace to
work around WiFi networks with bad roaming implementations
- vrf: Rework interaction with netfilter/conntrack
- fq_codel: implement L4S style ce_threshold_ect1 marking
- sch: Eliminate unnecessary RCU waits in mini_qdisc_pair_swap()
BPF:
- Add support for new btf kind BTF_KIND_TAG, arbitrary type tagging
as implemented in LLVM14
- Introduce bpf_get_branch_snapshot() to capture Last Branch Records
- Implement variadic trace_printk helper
- Add a new Bloomfilter map type
- Track <8-byte scalar spill and refill
- Access hw timestamp through BPF's __sk_buff
- Disallow unprivileged BPF by default
- Document BPF licensing
Netfilter:
- Introduce egress hook for looking at raw outgoing packets
- Allow matching on and modifying inner headers / payload data
- Add NFT_META_IFTYPE to match on the interface type either from
ingress or egress
Protocols:
- Multi-Path TCP:
- increase default max additional subflows to 2
- rework forward memory allocation
- add getsockopts: MPTCP_INFO, MPTCP_TCPINFO, MPTCP_SUBFLOW_ADDRS
- MCTP flow support allowing lower layer drivers to configure msg
muxing as needed
- Automatic Multicast Tunneling (AMT) driver based on RFC7450
- HSR support the redbox supervision frames (IEC-62439-3:2018)
- Support for the ip6ip6 encapsulation of IOAM
- Netlink interface for CAN-FD's Transmitter Delay Compensation
- Support SMC-Rv2 eliminating the current same-subnet restriction, by
exploiting the UDP encapsulation feature of RoCE adapters
- TLS: add SM4 GCM/CCM crypto support
- Bluetooth: initial support for link quality and audio/codec offload
Driver APIs:
- Add a batched interface for RX buffer allocation in AF_XDP buffer
pool
- ethtool: Add ability to control transceiver modules' power mode
- phy: Introduce supported interfaces bitmap to express MAC
capabilities and simplify PHY code
- Drop rtnl_lock from DSA .port_fdb_{add,del} callbacks
New drivers:
- WiFi driver for Realtek 8852AE 802.11ax devices (rtw89)
- Ethernet driver for ASIX AX88796C SPI device (x88796c)
Drivers:
- Broadcom PHYs
- support 72165, 7712 16nm PHYs
- support IDDQ-SR for additional power savings
- PHY support for QCA8081, QCA9561 PHYs
- NXP DPAA2: support for IRQ coalescing
- NXP Ethernet (enetc): support for software TCP segmentation
- Renesas Ethernet (ravb) - support DMAC and EMAC blocks of
Gigabit-capable IP found on RZ/G2L SoC
- Intel 100G Ethernet
- support for eswitch offload of TC/OvS flow API, including
offload of GRE, VxLAN, Geneve tunneling
- support application device queues - ability to assign Rx and Tx
queues to application threads
- PTP and PPS (pulse-per-second) extensions
- Broadcom Ethernet (bnxt)
- devlink health reporting and device reload extensions
- Mellanox Ethernet (mlx5)
- offload macvlan interfaces
- support HW offload of TC rules involving OVS internal ports
- support HW-GRO and header/data split
- support application device queues
- Marvell OcteonTx2:
- add XDP support for PF
- add PTP support for VF
- Qualcomm Ethernet switch (qca8k): support for QCA8328
- Realtek Ethernet DSA switch (rtl8366rb)
- support bridge offload
- support STP, fast aging, disabling address learning
- support for Realtek RTL8365MB-VC, a 4+1 port 10M/100M/1GE switch
- Mellanox Ethernet/IB switch (mlxsw)
- multi-level qdisc hierarchy offload (e.g. RED, prio and shaping)
- offload root TBF qdisc as port shaper
- support multiple routing interface MAC address prefixes
- support for IP-in-IP with IPv6 underlay
- MediaTek WiFi (mt76)
- mt7921 - ASPM, 6GHz, SDIO and testmode support
- mt7915 - LED and TWT support
- Qualcomm WiFi (ath11k)
- include channel rx and tx time in survey dump statistics
- support for 80P80 and 160 MHz bandwidths
- support channel 2 in 6 GHz band
- spectral scan support for QCN9074
- support for rx decapsulation offload (data frames in 802.3
format)
- Qualcomm phone SoC WiFi (wcn36xx)
- enable Idle Mode Power Save (IMPS) to reduce power consumption
during idle
- Bluetooth driver support for MediaTek MT7922 and MT7921
- Enable support for AOSP Bluetooth extension in Qualcomm WCN399x and
Realtek 8822C/8852A
- Microsoft vNIC driver (mana)
- support hibernation and kexec
- Google vNIC driver (gve)
- support for jumbo frames
- implement Rx page reuse
Refactor:
- Make all writes to netdev->dev_addr go thru helpers, so that we can
add this address to the address rbtree and handle the updates
- Various TCP cleanups and optimizations including improvements to
CPU cache use
- Simplify the gnet_stats, Qdisc stats' handling and remove
qdisc->running sequence counter
- Driver changes and API updates to address devlink locking
deficiencies"
* tag 'net-next-for-5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (2122 commits)
Revert "net: avoid double accounting for pure zerocopy skbs"
selftests: net: add arp_ndisc_evict_nocarrier
net: ndisc: introduce ndisc_evict_nocarrier sysctl parameter
net: arp: introduce arp_evict_nocarrier sysctl parameter
libbpf: Deprecate AF_XDP support
kbuild: Unify options for BTF generation for vmlinux and modules
selftests/bpf: Add a testcase for 64-bit bounds propagation issue.
bpf: Fix propagation of signed bounds from 64-bit min/max into 32-bit.
bpf: Fix propagation of bounds from 64-bit min/max into 32-bit and var_off.
net: vmxnet3: remove multiple false checks in vmxnet3_ethtool.c
net: avoid double accounting for pure zerocopy skbs
tcp: rename sk_wmem_free_skb
netdevsim: fix uninit value in nsim_drv_configure_vfs()
selftests/bpf: Fix also no-alu32 strobemeta selftest
bpf: Add missing map_delete_elem method to bloom filter map
selftests/bpf: Add bloom map success test for userspace calls
bpf: Add alignment padding for "map_extra" + consolidate holes
bpf: Bloom filter map naming fixups
selftests/bpf: Add test cases for struct_ops prog
bpf: Add dummy BPF STRUCT_OPS for test purpose
...
Pull overflow updates from Kees Cook:
"The end goal of the current buffer overflow detection work[0] is to
gain full compile-time and run-time coverage of all detectable buffer
overflows seen via array indexing or memcpy(), memmove(), and
memset(). The str*() family of functions already have full coverage.
While much of the work for these changes have been on-going for many
releases (i.e. 0-element and 1-element array replacements, as well as
avoiding false positives and fixing discovered overflows[1]), this
series contains the foundational elements of several related buffer
overflow detection improvements by providing new common helpers and
FORTIFY_SOURCE changes needed to gain the introspection required for
compiler visibility into array sizes. Also included are a handful of
already Acked instances using the helpers (or related clean-ups), with
many more waiting at the ready to be taken via subsystem-specific
trees[2].
The new helpers are:
- struct_group() for gaining struct member range introspection
- memset_after() and memset_startat() for clearing to the end of
structures
- DECLARE_FLEX_ARRAY() for using flex arrays in unions or alone in
structs
Also included is the beginning of the refactoring of FORTIFY_SOURCE to
support memcpy() introspection, fix missing and regressed coverage
under GCC, and to prepare to fix the currently broken Clang support.
Finishing this work is part of the larger series[0], but depends on
all the false positives and buffer overflow bug fixes to have landed
already and those that depend on this series to land.
As part of the FORTIFY_SOURCE refactoring, a set of both a
compile-time and run-time tests are added for FORTIFY_SOURCE and the
mem*()-family functions respectively. The compile time tests have
found a legitimate (though corner-case) bug[6] already.
Please note that the appearance of "panic" and "BUG" in the
FORTIFY_SOURCE refactoring are the result of relocating existing code,
and no new use of those code-paths are expected nor desired.
Finally, there are two tree-wide conversions for 0-element arrays and
flexible array unions to gain sane compiler introspection coverage
that result in no known object code differences.
After this series (and the changes that have now landed via netdev and
usb), we are very close to finally being able to build with
-Warray-bounds and -Wzero-length-bounds.
However, due corner cases in GCC[3] and Clang[4], I have not included
the last two patches that turn on these options, as I don't want to
introduce any known warnings to the build. Hopefully these can be
solved soon"
Link: https://lore.kernel.org/lkml/20210818060533.3569517-1-keescook@chromium.org/ [0]
Link: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/log/?qt=grep&q=FORTIFY_SOURCE [1]
Link: https://lore.kernel.org/lkml/202108220107.3E26FE6C9C@keescook/ [2]
Link: https://lore.kernel.org/lkml/3ab153ec-2798-da4c-f7b1-81b0ac8b0c5b@roeck-us.net/ [3]
Link: https://bugs.llvm.org/show_bug.cgi?id=51682 [4]
Link: https://lore.kernel.org/lkml/202109051257.29B29745C0@keescook/ [5]
Link: https://lore.kernel.org/lkml/20211020200039.170424-1-keescook@chromium.org/ [6]
* tag 'overflow-v5.16-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux: (30 commits)
fortify: strlen: Avoid shadowing previous locals
compiler-gcc.h: Define __SANITIZE_ADDRESS__ under hwaddress sanitizer
treewide: Replace 0-element memcpy() destinations with flexible arrays
treewide: Replace open-coded flex arrays in unions
stddef: Introduce DECLARE_FLEX_ARRAY() helper
btrfs: Use memset_startat() to clear end of struct
string.h: Introduce memset_startat() for wiping trailing members and padding
xfrm: Use memset_after() to clear padding
string.h: Introduce memset_after() for wiping trailing members/padding
lib: Introduce CONFIG_MEMCPY_KUNIT_TEST
fortify: Add compile-time FORTIFY_SOURCE tests
fortify: Allow strlen() and strnlen() to pass compile-time known lengths
fortify: Prepare to improve strnlen() and strlen() warnings
fortify: Fix dropped strcpy() compile-time write overflow check
fortify: Explicitly disable Clang support
fortify: Move remaining fortify helpers into fortify-string.h
lib/string: Move helper functions out of string.c
compiler_types.h: Remove __compiletime_object_size()
cm4000_cs: Use struct_group() to zero struct cm4000_dev region
can: flexcan: Use struct_group() to zero struct flexcan_regs regions
...
Firmware can trigger a missed beacon indication, this is not the same as a
lost signal.
Flag to Linux the missed beacon and let the WiFi stack decide for itself if
the link is up or down by sending its own probe to determine this.
We should only be signalling the link is lost when the firmware indicates
Fixes: 8e84c25821 ("wcn36xx: mac80211 driver for Qualcomm WCN3660/WCN3680 hardware")
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211027232529.657764-1-bryan.odonoghue@linaro.org
An SMD capture from the downstream prima driver on WCN3680B shows the
following command sequence for connected scans:
- init_scan_req
- start_scan_req, channel 1
- end_scan_req, channel 1
- start_scan_req, channel 2
- ...
- end_scan_req, channel 3
- finish_scan_req
- init_scan_req
- start_scan_req, channel 4
- ...
- end_scan_req, channel 6
- finish_scan_req
- ...
- end_scan_req, channel 165
- finish_scan_req
Upstream currently never calls wcn36xx_smd_end_scan, and in some cases[1]
still sends finish_scan_req twice in a row or before init_scan_req. A
typical connected scan looks like this:
- init_scan_req
- start_scan_req, channel 1
- finish_scan_req
- init_scan_req
- start_scan_req, channel 2
- ...
- start_scan_req, channel 165
- finish_scan_req
- finish_scan_req
This patch cleans up scanning so that init/finish and start/end are always
paired together and correctly nested.
- init_scan_req
- start_scan_req, channel 1
- end_scan_req, channel 1
- finish_scan_req
- init_scan_req
- start_scan_req, channel 2
- end_scan_req, channel 2
- ...
- start_scan_req, channel 165
- end_scan_req, channel 165
- finish_scan_req
Note that upstream will not do batching of 3 active-probe scans before
returning to the operating channel, and this patch does not change that.
To match downstream in this aspect, adjust IEEE80211_PROBE_DELAY and/or
the 125ms max off-channel time in ieee80211_scan_state_decision.
[1]: commit d195d7aac0 ("wcn36xx: Ensure finish scan is not requested
before start scan") addressed one case of finish_scan_req being sent
without a preceding init_scan_req (the case of the operating channel
coinciding with the first scan channel); two other cases are:
1) if SW scan is started and aborted immediately, without scanning any
channels, we send a finish_scan_req without ever sending init_scan_req,
and
2) as SW scan logic always returns us to the operating channel before
calling wcn36xx_sw_scan_complete, finish_scan_req is always sent twice
at the end of a SW scan
Fixes: 8e84c25821 ("wcn36xx: mac80211 driver for Qualcomm WCN3660/WCN3680 hardware")
Signed-off-by: Benjamin Li <benl@squareup.com>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211027170306.555535-4-benl@squareup.com
Without ieee80211_ops->flush implemented to empty HW queues, mac80211 will
do a 100ms dead wait after stopping SW queues, before leaving the operating
channel to resume a software connected scan[1].
(see ieee80211_scan_state_resume)
This wait is correctly included in the calculation for whether or not
we've exceeded max off-channel time, as it occurs after sending the null
frame with PS bit set. Thus, with 125 ms max off-channel time we only
have 25 ms of scan time, which technically isn't even enough to scan one
channel (although mac80211 always scans at least one channel per off-
channel window).
Moreover, for passive probes we end up spending at least 100 ms + 111 ms
(IEEE80211_PASSIVE_CHANNEL_TIME) "off-channel"[2], which exceeds the listen
interval of 200 ms that we provide in our association request frame. That's
technically out-of-spec.
[1]: Until recently, wcn36xx performed software (rather than FW-offloaded)
scanning when 5GHz channels are requested. This apparent limitation is now
resolved -- see commit 1395f8a6a4d5 ("wcn36xx: Enable hardware scan offload
for 5Ghz band").
[2]: in quotes because about 100 ms of it is still on-channel but with PS
set
Signed-off-by: Benjamin Li <benl@squareup.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211027170306.555535-3-benl@squareup.com
ATH10K chips are used it wide range of routers,
accesspoints, range extenders, network appliances.
On these embedded devices, calibration data is often
stored on the main system's flash and was out of reach
for the driver.
To bridge this gap, ath10k is getting extended to pull
the (pre-)calibration data through nvmem subsystem.
To do this, a nvmem-cell containing the information can
either be specified in the platform data or via device-tree.
Tested with:
Netgear EX6150v2 (IPQ4018 - pre-calibration method)
TP-Link Archer C7 v2 (QCA9880v2 - old calibration method)
Cc: Robert Marko <robimarko@gmail.com>
Cc: Thibaut VARÈNE <hacks@slashdirt.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211016234609.1568317-1-chunkeey@gmail.com
Commit 6f4d70308e ("ath11k: support SMPS configuration for 6 GHz") changed
"if (ht_cap & WMI_HT_CAP_DYNAMIC_SMPS)" to "if (ht_cap &
WMI_HT_CAP_DYNAMIC_SMPS || ar->supports_6ghz)" which means
NL80211_FEATURE_DYNAMIC_SMPS is enabled for all chips which support 6 GHz.
However, WCN6855 supports 6 GHz but it does not support feature
NL80211_FEATURE_DYNAMIC_SMPS, and this can lead to MU-MIMO test failures for
WCN6855.
Disable NL80211_FEATURE_DYNAMIC_SMPS for WCN6855 since its ht_cap does not
support WMI_HT_CAP_DYNAMIC_SMPS. Enable the feature only on QCN9074 as that's
the only other device supporting 6 GHz band.
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-01720.1-QCAHSPSWPL_V1_V2_SILICONZ_LITE-1
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210914163726.38604-3-jouni@codeaurora.org
Add the missing endpoint max-packet sanity check to probe() to avoid
division by zero in ath10k_usb_hif_tx_sg() in case a malicious device
has broken descriptors (or when doing descriptor fuzz testing).
Note that USB core will reject URBs submitted for endpoints with zero
wMaxPacketSize but that drivers doing packet-size calculations still
need to handle this (cf. commit 2548288b4f ("USB: Fix: Don't skip
endpoint descriptors with maxpacket=0")).
Fixes: 9cbee35868 ("ath6kl: add full USB support")
Cc: stable@vger.kernel.org # 3.5
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211027080819.6675-3-johan@kernel.org
Add the missing endpoint max-packet sanity check to probe() to avoid
division by zero in ath10k_usb_hif_tx_sg() in case a malicious device
has broken descriptors (or when doing descriptor fuzz testing).
Note that USB core will reject URBs submitted for endpoints with zero
wMaxPacketSize but that drivers doing packet-size calculations still
need to handle this (cf. commit 2548288b4f ("USB: Fix: Don't skip
endpoint descriptors with maxpacket=0")).
Fixes: 4db66499df ("ath10k: add initial USB support")
Cc: stable@vger.kernel.org # 4.14
Cc: Erik Stromdahl <erik.stromdahl@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211027080819.6675-2-johan@kernel.org
