| Commit message (Collapse) | Author | Age |
| ... | |
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Check for the value of sus_res_mcastbcast_filter_valid before resetting it
to false. The driver can unload/load while the wlan driver has been
suspended. In such a case, the value of flag sus_res_mcastbcast_filter
would be reset. Check for the value of the flag to take care of this case.
Change-Id: I64ea6f304d4f7535dc1fb47ca3cb0b87654a08ca
CRs-Fixed: 615781
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Break up the IsLoadUnloadInProgress flag to distinguish between load and
unload condition. The regulatory callback needs to be called during load
sequence, whereas during unload it needs to be avoided.
Change-Id: I465fa935f78084cd33597a6bef3e1dae5147059b
CRs-Fixed: 614579
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Pulling in below ROC related fixes from Pronto mainline,
599865, 599867, 566005
Change-Id: I8c16e7cef57f107c82e0aa3adaa50f5a9a1f11a1
CRs-Fixed: 613164
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Changes to set NOA and opportunistic power save parameters
to FW on receiving private IOCTL request from supplicant.
Change-Id: Ic8f1a2aaed659126c947dfb1d8b5296b3cbe09be
CRs-fixed: 612134
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
During the check to prevent the Host from configuring more than maximum
number of interfaces than that are allowed by the Firmware, the station
management entity is deferencing cross layer context information. This
fix provides an alternative solution to avoid this issue.
CRs-Fixed: 611191
Change-Id: I8a34bdb3389ce9d8240990e4f6e90b47cbfd0031
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In SAP mode the support for iwpriv commands set11NRates,
and set11ACRates features is added.
CRs-Fixed: 594430
Change-Id: I5e1674b256f97a9b0b8d81364c48df30b2cd80a3
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This patch separated setTxPower and setMaxPower two user ioctls apart.
Also, setTxPower is extended from SAP to station as well. Then, Rome has
no setMaxTxPower interface. A caching logic is implemented in wma layer to
assure that setTxPower's power value does not go beyond the power value
of setMaxPower.
CRs-Fixed: 594430
Change-Id: I0692a2b0af68f36adad307092841e738f9c01a04
|
| | |\ \ \ \ \ |
|
| | |/| | | | |
| | |/ / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Release 1.0.0.35 QCACLD WLAN Driver
* origin/caf/caf-wlan/master:
Cafstaging Release 1.0.0.35
wlan: Use the session's rate_flag for finding HT/VHT's MCS index
wlan: fix the set mcast rate issue
wlan:qcacld:rome_ibss:tune peer departure time out values
qcacld: Add SSR crash shutdown API
qcacld: Add support to dump copy engine debug registers
qcacld: Adding VOS_BUG after detecting HTC header corruption
qcacld: cache user configured patterns in WMA without corruption
qcacld : Adding logs for CSA/PNO code flow
WMA: Changes to use adf_os_mem_alloc in interrupt context
qcacld: CL 837998 - update fw common interface files
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Cafstaging Release 1.0.0.35
Change-Id: I77912b609bb6ef7d65bfa7288eb8a51797742097
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
The short_gi bit of rate_flag is used to choose HT/VHT's rate table
CRs-Fixed: 584490
Change-Id: I0cfde58497e379a165720199fba64dccc4197a29
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Firmware mcast implementation is decoupled from ucast. The host also needs
to decouple the ucast and mcast. Also, the MCS table search needs to
consider both short gate flag and band width enum. Then, the application
interface has a change to allow the application to specify the NSS setting.
CRs-Fixed: 606821
Change-Id: I046bf74a126faf9141014a3e227b846a9d0cbeea
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
With probe request at FW side, false peer departure is still
seen if we use constant time out value (4 secs).
So we tune the values such that they slowly increase based on
the number of connected peers
Change-Id: I031e3c15ef75537512ddb3701e6303f0fa146898
CRs-Fixed: 614441
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Add the SSR crash shutdown API in order to support collect target RAM
dump when kernel panic happens or other subsystems crash.
Change-Id: Ibe97baee5e1d782ecdc173a2f8d4dc9e841a2a37
CRs-Fixed: 609070
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
When target assert happens, sometimes host driver fails to access target
RAM area to collect RAM dump because of copy engine failure. Add support
to print copy engine debug registers in order to debug this case.
Change-Id: I1305db88fc754e898054a823682f4aea344a7fa6
CRs-Fixed: 613150
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Adding VOS_BUG to force device into download mode to get APSS
and FW RAMDUMP for further analysis of HTC header corruption
issue
Change-Id: Id207a5e498f526ac68fde1a31ddfa32c7da5503e
CRs-Fixed: 614665
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
User configured patterns using iwpriv are getting corrupted
when caching in WMA due to wrong pointer dereferenced.
CRs-Fixed: 614028
Change-Id: Ib903c341a311f4aa0105099f44c4a570c160400a
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Adding some missing logs to help in debugging.
Change-Id: Ib7dff0e98fcd5619515ed6622b0acfd9c7e40160
CRs-fixed: 614778
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Replace vos_mem_alloc with adf_os_mem_alloc
in interrupt conext to allocate memory
Change-Id: Ia413964dd4746ee81892287e9ab5c82b46d8261c
CRs-fixed: 614898
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Add event id and typedefs for L1SS statistics in wmi header files.
Change-Id: I9d5ea0fcaee317d2f0e6dc6326004c088027cfd5
CRs-Fixed: 614079
|
| | |\ \ \ \ \ |
|
| | |/| | | | |
| | |/ / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Release 1.0.0.34 QCACLD WLAN Driver
* origin/caf/caf-wlan/master:
Cafstaging Release 1.0.0.34
qcacld: ipa: Use adf_os_mem_XXX APIs
wlan: cld P2P ping fail, regression revert.
qcacld:Fix the double memory free in scan start command
qcacld: Selectively build p2p interface bring up per platform
qcacld: Fix of 1 byte memory corruption during driver unload
qcacld: Increasing hostapd driver command buffer size check
wlan: make target failure dump 64 bit compatible
qcacld: Fix tbtt adjustment on beacon updation
wlan: corrected bitwise operation in get intf mac addr
qcacld: Fix for schedule scan print
qcacld: Fix of memory alloc failure in vdev detach
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Cafstaging Release 1.0.0.34
Change-Id: I5abc2cca8f7143bdf69c215d534f7591ec3d5c58
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Use adf_os_mem_XXX APIs instead of vos_mem_XXX APIs.
adf_os_mem_XXX APIs are robust to use for IPA implementation.
There are issues with IPA when vos_mem_malloc() was used because
not doing memzero but adf_os_mem_alloc() takes care of memzero.
For consistency purpose all vos_mem_XXX APIs are replaced with
adf_os_mem_XXX APIs.
Change-Id: I939caa3ddfa04125e818db2f793828195920be1b
CRs-fixed: 614449
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Revert change made regression for P2P.
https://review-android.quicinc.com/#/c/620743
Change-Id: I02639a97f515b61e411376085ec5cc666a273a44
CRs-fixed: 614064
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
wma doesn't allow back to back scan. Here scan completion timer is
getting started after wma sends the start scan command. But if already
scan is running then scan completion timer start will fail. If it
fails currently we are freeing the scan buffer memory which is
currently getting used in other layer and the same buffer will get freed
again upon getting scan complete event. Recieve buffer also got the same
memory. Because of double free in tx path system crashed while rx data is
recieved because of access to already freed buffer.
So scan completion timer is started even before we send the scan start
command so that we can undo the timer operation incase any of the
operation fails.
Change-Id: Icfc351a79639906852083c8be3117f568d6065d7
CRs-Fixed: 613209
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
To bring up P2P interface selectively per platform,
added a new build flag WLAN_OPEN_P2P_INTERFACE.
Change-Id: Ia2b77d672eee115c67f44bb8f7fa9e1ee4cb0864
CRs-fixed: 594777
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Following is the crash scenario:
1) vos_close() is calling wma_close() which is freeing wmi_handle
2) After freeing wmi_handle vos_close() is calling HTCStop()
3) HTCStop() which is calling HIFStop and HIStop() is calling
hif_buffer_cleanup() to flush send buffer commands using
hif_send_buffer_cleanup_on_pipe()
4) And now while flushing HIF buffer commands it calls
wmi_htc_tx_complete() which de-references already freed
wmi_handle pointer and decrements it casuing a single byte of
memory corruption.
5) Fixing this issue by moving down wmi_handle free after HTCStop()
during driver unload such that when wmi_htc_tx_complete()
accesses wmi_handle while flushing HIF send buffer it gets
valid wmi_handle.
CRs-Fixed: 613863
Change-Id: I4f053d930f797b7142479897a2238fcad438e023
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In batch scan driver command string buffer length is coming as
4096 so increasing driver command string buffer size check in
hostapd dev ioctl handler
Change-Id: Ie31752ef2622cfdfc6566a34c0c1ec53936e455e
CRs-Fixed: 613978
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
32 bit target addresses were processed as a pointer on the host.
On a 64 bit host, this leads to a compiler warning. Using the read
addresses directly without the unnecessary typecast on the host.
Change-Id: I26e597d806ba2233bd4c04ae8497f46dc6127d74
CRs-fixed: 613272
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Whenever firmware updates TBTT offset adjustment through
WMI_TBTTOFFSET_UPDATE_EVENTID, the negative of the TBTT adjustment
offset is calculated and updated in timestamp of beacon buffer.
This beacon timestamp updation is not handled whenever UMAC is
sending updated beacon template. This is causing imbalance NOA
syncup in MCC scenario. Fix that.
Change-Id: I73f27bf78b30f1daaeb3c80954303e0f608b2927
CRs-Fixed: 593601
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In wlan_hdd_get_intf_addr function corrected
logic for checking unused intf mac address.
Change-Id: Ida7ed7a3ba25ae327e868b740c14122252495b16
CRs-Fixed: 609725
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
wpa_supplicant periodically scans neighboring access points if the
wiphy capability WIPHY_FLAG_SUPPORTS_SCHED_SCAN is set. Concurrent
sessions do not allow scheduled scan. The log level of the print is
lowered to info. Also corrected another print on flushing pmksa
entries.
Change-Id: Iaf0d83c21e7197a230552286f2c40b6161bf6d4b
CRs-Fixed: 610942
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
wma_vdev_detach() isusing vos_mem_malloc() in atomic context
to allocate memory. Chaning vsv_mem_malloc() to adf_os_mem_alloc()
such that wma_vdev_detach() can allocate memory in atomic context.
Change-Id: Iccfef9d115b5035540aa75d6630cda1f7035f71a
CRs-Fixed: 613766
|
| | |\ \ \ \ \ |
|
| | |/| | | | |
| | |/ / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Release 1.0.0.33 QCACLD WLAN Driver
* origin/caf/caf-wlan/master:
cafstaging release 1.0.0.33
qcacld: Cleanup HIF PCI suspend/resume
qcacld/hif: Add debug log after potential hw_index corruption
wlan: Fix format specifier in print
qcacld: Featurize self recovery
wlan: fix format specifier in print statement
qcacld: Fix of memory leak in LIM DEL BSS RSP handler
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
cafstaging Release 1.0.0.33
Change-Id: Iebfa564b1b88b50002733cd4494de13ec56ab304
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Remove the function calls from HIF suspend and resume functions
to HDD suspend/resume functions that are no longer being called
from PCI suspend/resume under the disabled flag -
WLAN_LINK_UMAC_SUSPEND_WITH_BUS_SUSPEND.
They are now called from cfg80211 suspend and resume functions
respectively.
Change-Id: I462a5cb3e2ce4d789a8df6fca0db73627439f3c0
CRs-Fixed: 613659
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
This issue happens when doing send complete of wmi
command during the device suspend case. In normal
condition, we should never run out of hif free
buffers used for tx/rx completion. A potential
issue here is that the hw_index might be wrong during
suspend which leads to a busy loop trying to do
bogus send completex . Add some debug information
when we run out of hif send completion buffer, from
this debug log we could find the signature and come
up with a WAR in a separate patch when we see the issue
next time.
Change-Id: Ica78af10376156be5adde08523f13f08581df6bd
CRs-Fixed: 612173
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Use correct format specifier to fix compiler warning for a 64 bit compiler
Change-Id: If9f132a61781e4032e4ac43b777db3ba1f4b9c82
CRs-fixed: 613266
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Featurize the self recovery; and it should be disabled in all internal
builds.
Change-Id: I98fe2d883b26a83369979c69e4fbe2433b42c504
CRs-Fixed: 613122
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Use correct format specifier to fix compiler warning
Change-Id: I97babc66d12c2fe8d12e8a5658f7a5902df3bb7a
CRs-fixed: 613262
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Fixing limHandleDeleteBssRsp() to free pDelBss pointer object
allocated by WMA when LIM fails to find session ID
Change-Id: Iae865172dbea882254899c997ad11a1e03d31d51
CRs-Fixed: 613312
|
| | |\ \ \ \ \ |
|
| | | |\| | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Release 1.0.0.32 QCACLD WLAN Driver
* origin/caf/caf-wlan/master:
Cafstaging Release 1.0.0.32
qcacld: Fixing crashes caused by dangling vdev pointer
wlan: MAC Address Logging Fix.
qca-cld: translation of CoEx frequency range reportings
Kbuild: add LTE CoEx feature dependency to platform driver
qcacld: Fix of kernel panic in PCIe interrupt handler
qcacld: ipa: Modify IPA header file path
wlan: make cfg.dat file read 64 bit compatible
wlan: Multiple TX PAUSE type event handle
wlan: TL: removing newlines from the debug logs
wlan: WDA: removing newlines from the debug logs
wlan: VOSS: removing newlines from the debug logs
wlan: MAC Address Logging Fix.
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Cafstaging Release 1.0.0.32
Change-Id: Id0d53b49b83fe3717147ea3c3cc25590da006044
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
wma_vdev_deatch() should not call ol_txrx_vdev_deatch() for
same vdev object otherwise it can cause access of same vdev
object which is already freed and crash because of that. Adding
a wma_vdev_detach spin lock to protect ol_txrx_vdev_detach()
getting called 2 or more times for same vdev from different threads.
Change-Id: I005b388c188b0885516fc1d3bd2e9b2722627481
CRs-Fixed: 612606
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Use pre-defined macro's to log MAC addresses.
Change-Id: I1884ad5ee049a247b51a35a9a6687f318585aeae
CRs-Fixed: 573100
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Add boundary to center frequency translation in WMA, and reverted
the previous change in HDD for the same.
Change-Id: I5ee39993fe549ea68167cd72e2ecbcbefd516ae2
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Add the dependency of LTE CoEx feature inclusion to platform driver
in the Kbuild file.
Change-Id: I7e9138971164d0973064ba0ddbb8bc31d8822e0c
|