| Age | Commit message (Collapse) | Author |
|---|
|
Now that all the needed pieces are in place implement suspend support QCA6390
PCI devices. All other devices will return -EOPNOTSUPP during suspend. The
suspend is implemented by switching the firmware to WoW mode during suspend, so
the firmware will be running on low power mode while host is in suspend.
At the moment we are not able to shutdown and fully power off the device due to
bugs in MHI subsystem, so WoW mode is a workaround for the time being.
During suspend we enable WoW mode, disable CE irq and DP irq, then put MHI to
suspend state. During resume, driver resumes MHI firstly, then enables CE irq
and dp IRQ, and sends WoW wakeup command to firmware.
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/1607708150-21066-11-git-send-email-kvalo@codeaurora.org
|
|
Stop dp rx pktlog when entering suspend and reap the mon_status buffer to keep
it empty. During resume restart the reap timer.
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/1607708150-21066-7-git-send-email-kvalo@codeaurora.org
|
|
Found by latest checkpatch.
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1603803705-22447-1-git-send-email-kvalo@codeaurora.org
|
|
Start a timer to update REO HP if HP isn't updated to target.
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/1601544890-13450-6-git-send-email-kvalo@codeaurora.org
|
|
The timer is to check if TCL HP isn't updated to target.
The timer will postpone itself if there are TX operations
during the interval, otherwise the timer handler updates
the HP again so the index value in HP register will be
forwarded to target register, and the timer stops afterwards.
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/1601544890-13450-5-git-send-email-kvalo@codeaurora.org
|
|
Add packet log support for QCA6390, otherwise the data connection will stall
within a minute or so. Enable it via debugfs and use trace-cmd to capture the
pktlogs.
echo 0xffff 1 > /sys/kernel/debug/ath11k/qca6390\ hw2.0/mac0/pktlog_filter
The mon status ring doesn't support interrupt so far, so host starts
a timer to reap this ring. The timer handler also reaps the
rxdma_err_dst_ring in case of monitor mode.
As QCA6390 requires bss created ahead of starting vdev, so check
vdev_start_delay for monitor mode.
For QCA6390, it uses wbm_desc_rel_ring to return descriptors.
It also uses rx_refill_buf_ring to fill mon buffer instead of
rxdma_mon_buf_ring.
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/1601463073-12106-2-git-send-email-kvalo@codeaurora.org
|
|
For QCA6390, only one pdev is created and this pdev manages both lmacs, thus
both rxdmas. So host needs to initialize all rxdma related rings for one pdev.
Another difference is for QCA6390, host fills rxbuf to firmware and firmware
further fills the rxbuf to rxbuf ring for each rxdma.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.1.0.1-01238-QCAHKSWPL_SILICONZ-2
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597555891-26112-4-git-send-email-kvalo@codeaurora.org
|
|
For QCA6390, it uses peer_map and peer_unmap V1. IPQ8074 uses V2.
Redefine previous definition to peer_map2 and peer_unmap2.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.1.0.1-01238-QCAHKSWPL_SILICONZ-2
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1597576599-8857-13-git-send-email-kvalo@codeaurora.org
|
|
Add support for collecting and dumping the ring backpressure
stats via debugfs. Stats are dumped only if events are
received for the specific ring.
Below command can be used to obtain these stats as part of soc dp stats.
cat /sys/kernel/debug/ath11k/ipq8074/soc_dp_stats
Sample Output - When No stats available:
Backpressure Stats
==================
No Ring Backpressure stats received
Sample Output - When ring bp stats available for specific ring
Backpressure Stats
==================
Ring: REO2SW1_RING
count: 1
hp: 2
tp: 2
seen before: 4ms
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.1.0.1-01213-QCAHKSWPL_SILICONZ-1
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1591768308-32005-3-git-send-email-srirrama@codeaurora.org
|
|
HTT EXT stats comes in stream of TLVs spanning over multiple
messages. Currently completion is being sent for each message
which is creating a race where stats_req is being accessed
for filling in second message after the memory is already
freed in release operation. Fix this by issuing completion
once all the messages are received and processed. Driver
knows this info from DONE bit set in htt msg.
Also fix locking required for htt stats.
Co-developed-by: Miles Hu <milehu@codeaurora.org>
Signed-off-by: Miles Hu <milehu@codeaurora.org>
Signed-off-by: Pradeep Kumar Chitrapu <pradeepc@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1589221074-28778-1-git-send-email-pradeepc@codeaurora.org
|
|
we are sending the reo flush command for the deleted peer
tid after the ageout period reaches 1 second. This handling
causes reo ring get full when more than 128 clients are
disconnected continuously. so added the count for flush list
and reo flush command is triggered after the list count reaches
the threshold value, it is configured as 64 (half of the reo ring).
This will avoid the situation where reo ring get full.
Signed-off-by: Karthikeyan Periyasamy <periyasa@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1587552378-4884-1-git-send-email-periyasa@codeaurora.org
|
|
Increase the tx completion ring size to 0x8000.Also set the
idr size to be same as the completion ring size.
This avoids backpressure on the TX Completion and
corresponding TCL Data ring during high data traffic.
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1585567006-9173-1-git-send-email-srirrama@codeaurora.org
|
|
As Hash based reo destination selection is configured,
the decapped packets reach different reo destintion rings
based on the destintaion ring selected for the computed hash (based on
the 5-tuple {ip src/ip dst/src port/dst port/protocol}) by hw and
as configured by driver.
Hence the current implementation of amsdu list based processing after all
the subframes of amsdu are received (since all msdu's for a pdev are
received in same reo dest ring), is not applicable here and hence is
replaced with per msdu based handling as these subframes
can be received in different reo dest rings.
Also, as some of the rx descriptor fields might be valid only for the
first msdu (for ex. received 80211 header, encryption type, etc),
it might not be useful now as we cannot sync between different
subframes received in different rings. Hence do not rely on those
fields and replace them with fieds valid only on per msdu descriptors.
Also cache other details such as encryption type for a peer so that
it can be reused when a packet is received from it.
Co-developed-by: Tamizh Chelvam Raja <tamizhr@codeaurora.org>
Signed-off-by: Tamizh Chelvam Raja <tamizhr@codeaurora.org>
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
Current implementation of pdev based reo destination ring
selection is replaced by hash based ring selection so as to
ensure all the available rings are utilized for better performance.
The 4 reo destination rings are selected by the HW based on the
hash value computed from the received packet based on the 5 tuple
{ip src/ip dst/src port/dst port/protocol}. Out of the 32 hash values
used by the hw, the driver assigns 8 values per reo destination ring
to each of the 4 reo destination rings.
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
The Firmware sends HTT event to host whenever there is a
backpressure on RX rings, Handling such event and dumping
info on the console under the "ATH11K_DBG_DP_HTT" debug level.
Fetching RX ring backpressure histogram from FW via htt_stats debugfs.
#echo "24" > /sys/kernel/debug/ath11k/ipq8074/macX/htt_stats_type
#cat /sys/kernel/debug/ath11k/ipq8074/macX/htt_stats
Signed-off-by: Vikas Patel <vikpatel@codeaurora.org>
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
IPQ8074 HW has support to verify the PN of the received frames.
For all frames except for fragmented ones, HW checks the PN and
delivers them to the driver. For fragmented frames, driver is
required to do a little more; it has to reassemble the fragments
and then reinject them to the HW for verifying the PN. Currently,
to keep the logic simple, PN verifcation is disabled in HW and is
handled in mac80211 for all the frames (fragmented and unfragmented).
On the contrary, offloading PN Validation to the HW brings important
benefits. It reduces CPU cycles spent on the host CPU for verifying
the same; helps in enabling features which improve performance like
mac80211 fast RX path, enabling multiple REO rings for parallel RX
processing, 802.11 decapsulation offloading. All these features are
dependent on PN offload which in turn is dependent on handling of
the received fragments in the driver.
When TKIP security is used, additional handling is required while
processing the fragments; since MIC is computed on an MSDU in TKIP,
only the last fragment has the MIC info. In this case, driver has to
compute the MIC after reassembly and compare it against the MIC
present in the frame. For this, MICHAEL_MIC kernel crypto library
APIs are used and the dependencies are appropriately set.
Signed-off-by: Manikanta Pubbisetty <mpubbise@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
Currently rx tid setup is happening for TID 0 and TID 16
during peer setup. And if other TID packets received for
the peer it will be redirected to rx error ring and not through
reo ring. And this rx tid configuration cannot be done
in the rx error ring path since it is a atomic context.
So moving the rx tid setup for all tids during the peer setup.
This is required to enable PN offload functionality to route
all packets through reo ring.
Co-developed-by: Tamizh Chelvam <tamizhr@codeaurora.org>
Signed-off-by: Tamizh Chelvam <tamizhr@codeaurora.org>
Signed-off-by: Govindaraj Saminathan <gsamin@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
Parse and store the out-of-band rates reported by the FW.
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
This patch enables HE tlvs in ppdu stats for pktlog lite mode.
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Anilkumar Kolli <akolli@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
the current code does 4 memcpys for each completion frame.
1) duplicate the desc
2 + 3) inside kfifo insertion
4) kfifo remove
The code simply drops the kfifo and uses a trivial ring buffer. This
requires a single memcpy for insertion. There is no removal needed as
we can simply use the inserted data for processing. As the code runs
inside the NAPI context it is atomic and there is no need for most of
the locking.
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
Firmware expects all the required REO destination rings setup
while processing wmi_init command. Not doing this causes connected
stations getting disconnected and not able to connect back.
Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
Currently HTT_PPDU_STATS_TAG_USR_COMPLTN_ACK_BA_STATUS tag of PPDU stats
doesn't have valid success bytes info of management frames.
So skip update peer stats for management packets.
Signed-off-by: Venkateswara Naralasetty <vnaralas@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
remove unused counter to avoid taking locks inorder to optimize
cpu load.
Using Flamegraph, cpu usage of ath11k_dp_tx() observed to be decreased
from 5.58% to 3.74% with iperf traffic running with 80MHz bandwidth ap
mode.
Signed-off-by: Pradeep Kumar Chitrapu <pradeepc@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
The Previous configuartion of htt stats sounding info and pdev cca stats
are invalid due to that getting time out error.
Changing htt stats sounding info value from 0xFF to 0x00 and htt pdev cca
stats from 0x10 to 0x00
Signed-off-by: Ganesh Sesetti <gseset@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
|
|
ath11k is a new driver for Qualcomm IEEE 802.11ax devices, first
supporting only IPQ8074 SoC using the shared memory AHB bus. ath11k
uses mac80211 and supports AP, Station and Mesh modes.
Even though ath11k has some similar code as with ath10k (especially
the WMI layer) it was concluded to be simpler to have a "clean start"
for ath11k code base and not try to share the code with ath10k. This
makes maintenance easier and avoids major changes in ath10k, which
would have significantly increased the risk of regressions in existing
setups.
Even though the driver is very similar with ath10k but there are major
differences as well. The datapath is completely different. ath11k
supports multiple MACs, called "soc" in the firmware interface. And
there's only one WMI interface to support.
Currently ath11k supports only IEEE 802.11ac mode, but patches for
802.11ax are available and they will be submitted after ath11k is
accepted to upstream.
The firmware images are available from ath11k-firmware repository but
they will be also submitted to linux-firmware:
https://github.com/kvalo/ath11k-firmware
This was tested with firmware version WLAN.HK.2.1.0.1-00629-QCAHKSWPL_SILICONZ-1.
The driver has had multiple authors who are listed in alphabetical
order below.
Signed-off-by: Anilkumar Kolli <akolli@codeaurora.org>
Signed-off-by: Bhagavathi Perumal S <bperumal@codeaurora.org>
Signed-off-by: Ganesh Sesetti <gseset@codeaurora.org>
Signed-off-by: Govindaraj Saminathan <gsamin@codeaurora.org>
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Karthikeyan Periyasamy <periyasa@codeaurora.org>
Signed-off-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Maharaja Kennadyrajan <mkenna@codeaurora.org>
Signed-off-by: Manikanta Pubbisetty <mpubbise@codeaurora.org>
Signed-off-by: Miles Hu <milehu@codeaurora.org>
Signed-off-by: Muna Sinada <msinada@codeaurora.org>
Signed-off-by: Pradeep Kumar Chitrapu <pradeepc@codeaurora.org>
Signed-off-by: Rajkumar Manoharan <rmanohar@codeaurora.org>
Signed-off-by: Sathishkumar Muruganandam <murugana@codeaurora.org>
Signed-off-by: Shashidhar Lakkavalli <slakkavalli@datto.com>
Signed-off-by: Sriram R <srirrama@codeaurora.org>
Signed-off-by: Sven Eckelmann <seckelmann@datto.com>
Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@codeaurora.org>
Signed-off-by: Venkateswara Naralasetty <vnaralas@codeaurora.org>
|
