Age | Commit message (Collapse) | Author |
|
QCA_WLAN_VENDOR_ATTR_CONFIG_LATENCY_LEVEL is supposed to carry u16
value in the request. Modify their type accordingly.
Bug: 222044962
Test: VTS tests
Change-Id: I7f5da5be31845ce214ef283279ec74d41474433b
|
|
Implement wifi_set_dtim_config API to send DTIM
configuration to driver/firmware.
Bug: 215653020
Test: DTIM configuration supported
Change-Id: I553447181c69e90d0f69134792ac9fc7c7ad7d8e
CRs-Fixed: 3118522
|
|
Increase MAX allowed number of whitelist SSIDs count to 8
Bug: 215795211
Test: Get roaming capabilities correctly on all related projects
Change-Id: I6ea8ff4f1623607b0180b243e230675f9ff7935c
CRs-Fixed: 2900468
|
|
Per WiFi HAL design, roaming feature should have nothing to do with
Gscan, but in current impl, these two are tightly coupled:
- Roaming capabilities is queried by vendor cmd GSCAN_GET_CAPABILITY.
- Set blacklist bssid returns error if Gscan not supported
So de-couple roaming feature with Gscan by:
- Set default roaming capabilities if Gscan not supported.
- Set blacklist bssid returns error if Roaming not supported.
Bug: 215795211
Test: Get roaming capabilities correctly on all related projects
CRs-Fixed: 2405118
Change-Id: I9f96546538dd4d07a1f930f107c0024436907927
|
|
Host driver does not use the attributes DRIVER_FW_LOCAL_WAKE_CNT_SZ
and CMD_EVENT_WAKE_CNT_SZ.
Also, driver unconditionally sets all of the LEGACY attributes to 0
in the response since they are not supported.
Hence, removing the relevant code.
Bug: 215804926
Test: Device and legacy pixels can get wake reason stats successfully
Change-Id: I63050a5c0645a32e4740342d3e84bd907394c2d7
CRs-Fixed: 2783515
|
|
Overflow may happen if request id is greater than max range of u8
type when trying to request driver/firmware version. Fix the issue by
not allowing request Id cross max range of u8 when trying to put with u8 api.
Bug: 213108298
Test: Device can get drirver version and firmware version
CRs-Fixed: 3107212
Change-Id: Ia42d9280d81f7019af025c4d5e7d4cb2e5ed916e
|
|
QCA_WLAN_VENDOR_ATTR_WIFI_INFO_DRIVER_VERSION and
QCA_WLAN_VENDOR_ATTR_WIFI_INFO_FIRMWARE_VERSION are supposed to carry u8
value in the request. Modify their type accordingly.
Bug: 213108298
Test: Device can get drirver version and firmware version
Change-Id: I78e4c1feeca3fcc3649299c5ad3a6972bf9bd60f
CRs-Fixed: 2755700
|
|
Validate the per packet TX rate codes and calculate
transmit rate only for valid rate codes.
Bug: 201226143
Test: Regression tests
Change-Id: I287d19d97970523b0d695e7dcbd606a49fa23f7c
CRs-Fixed: 2469251
|
|
llstats peerinfo has varying wifi rate stats. Since
llstats is a linear buffer with interface and peer
stats, we should include wifi rate stats length
while computing offset of next peer
Bug: 191419302
Test: Regression Test
Change-Id: Icc0395cf290268f4c2c96031389fae36ac48d47c
|
|
Wifihal uses mNumRadio from QCA_WLAN_VENDOR_ATTR_LL_STATS_NUM_RADIOS,
but it shall use mNumRadioAllocated which is based on the real radio
stats received.
This avoids memory out-of-bound access.
Bug: 190878523
Test: Regression test
CRs-Fixed: 2863682
Change-Id: Ia1c8484307182a3fab8fbfbde2a9be9047eade61
|
|
|
|
In Peer type message, the length of memcpy depends on
multiple data which can cause overflow due to corruption.
Added check for the length variable to properly perform
memcpy without overflow.
Bug: 183447882
Change-Id: I0e641a09af920e5d3db6fb3f2f526533907f74f9
CRs-Fixed: 2907367
|
|
Check the value of num_rx_discovery_attr before accessing
in NanMatch event to avoid out-of-bound write.
Bug: 184564368
CRs-Fixed: 2836431
Change-Id: I1a870c5f89394e2f66b1bbc16ea651ef5259772d
|
|
This commit fixes the mapping for HEAD scenarios by combining all of
them in the same table, this matches previous years mapping.
Bug: 162191325
Test: Make sure scenarios are mapped to the right table.
Change-Id: I23761becd615c11bde7cd076026ecedf54249d92
|
|
This commit maps the new SAR scenarios to the existing tables
Bug: 162191325
Test: Proper detection of SAR scenarios and tables
Change-Id: I273a6271f62c3cc81c40404980b72a88e0b15388
|
|
With support of dynamic interface create/remove 'commit : 5780336d',
wifi interface details shall be refreshed to acoomodate change in
interface list.
Bug: 181828707
Test: Manual Test - force to create virtual interface to make sure Hal
works normally.
Change-Id: I35ae3260688dfddaae864676a7f45ce24d4d10a9
Signed-off-by: Purushottam Kushwaha <quic_pkushwah@quicinc.com>
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/hardware/qcom/wlan/+/12835122
MUST ONLY BE SUBMITTED BY AUTOMERGER
Change-Id: I057b7e2fdc198e6cd9fd8c83fe7b19fe3c1beb41
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/hardware/qcom/wlan/+/12835123
MUST ONLY BE SUBMITTED BY AUTOMERGER
Change-Id: Ie1124542b065f69d434944cbc903b1e508f2f98a
|
|
This bring qca-vendor.h definitions as on 2020-09-03.
Bug: 160239223
CRs-Fixed: 2766035
Change-Id: Ica9a871bef2ab0b64c27df0cb85ae3ce66dddd98
|
|
Add support to parse channel RX time using QCA vendor interface of
LL stats, if supported by driver and report accurate CCA busy time.
Currently the driver/firmware indicates CCA busy time which
includes own RX time also and as such, CCA busy time reported by
firmware is not accurate.
Report accurate cca busy time to upper layers by subtracting self
RX time reported by firmware.
Bug: 160239223
Change-Id: I1218fce828b082d7145376e9fe7111c0e7bc5953
CRs-Fixed: 2766035
|
|
Added SPDX-license-identifier-Apache-2.0 SPDX-license-identifier-BSD to:
qcwcn/wifi_hal/Android.mk
Added SPDX-license-identifier-BSD to:
cld80211-lib/Android.mk
qcwcn/wcnss-service/Android.mk
qcwcn/wpa_supplicant_8_lib/Android.mk
Bug: 68860345
Bug: 151177513
Bug: 151953481
Test: m all
Exempt-From-Owner-Approval: janitorial work
Change-Id: I96fbf1a03a023509049ad11765b6860ab344499f
|
|
Currently, the driver control param file is created write-only,
so fix the check:
https://android.googlesource.com/kernel/msm-modules/qcacld/+/refs/heads/android-msm-coral-4.14-android11-qpr1/core/hdd/src/wlan_hdd_main.c#190
More importantly the open below uses O_WRONLY as only a write is
necessary, there is no need to check for read access.
Otherwise, WiFi fails to be enabled:
WifiHAL : Failed to access driver state control param Permission denied, 13 at /sys/kernel/boot_wlan/boot_wlan
Fixes: 8996467be9 ("wifihal: Update driver control param to check driver is ready")
Test: Manual build and functionality test
Signed-off-by: Pablo Mendez Hernandez <pablomh@gmail.com>
Signed-off-by: Marijn Suijten <marijn.suijten@somainline.org>
Change-Id: If6308cd790c69d2d57815ef6815c779258005eff
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/hardware/qcom/wlan/+/12124604
Change-Id: I66a07245e99c42da53a0bd6d2a0c8baf986fa535
|
|
Write driver control param and wait for successful update to make sure
driver is ready in LONU model.
Bug: 157090295
Test: Basic function test, SSR test
Change-Id: Ia1a690305266baafabe4c2557723f464ca858db1
|
|
Change-Id: I8896ee241761252ebde04f5d977f1a51a7892b41
|
|
Change-Id: I4ac4d460106ac3b242e9bfd714e1dedb5b9bbd67
|
|
The function wifi_add_or_remove_iface was taking care for
creation and deletion of virtual Iface. Added new separate
functions that will perform create and delete operations
instead of one single API.
Bug: 153852667
Bug: 154279995
Test: Pass wifi aware tests of CtsVerifier.
CRs-Fixed: 2633140
Change-Id: I3c175b1f5a59c6d022b2db51cfc72b48582ec3cf
|
|
This change added interface name start from "wifi" as part of
valid wifi interface name.
Bug: 153852667
Bug: 154279995
Test: Pass wifi aware tests of CtsVerifier.
CRs-Fixed: 2593436
Change-Id: Ic8154616c54905969e55b89c0a8f5cde923e5136
|
|
This is part of a cleanup to fix instances of the
-Wunreachable-code-loop-increment warning, which we plan to enable under
-Werror by default.
This isn't intended to be a functional change.
Bug: 150166387
Test: TreeHugger
Change-Id: Id7c2eea858bd8a1b6eb96d16f25bc00a6dc3e8b7
|
|
Instead of hardcoding base wlan interface as wlan0 get base wlan
interface index from hal_info
Bug: 140311866
Change-Id: I8f66207dfbcb791e02faa57a19460a4cedeaeca8
|
|
Bug: 140311866
Change-Id: If376a9e40ab3a1ad948cd95b15f30ba42fcb19bb
|
|
Update the WiFi-HAl as per the new interface definitions as needed.
Bug: 140311866
Test: Wifi Regression
Change-Id: I64b141468bd3a194d06e875ed4d3b782f0c9a4a3
|
|
Change-Id: I16a306e40c81663cd8311f1d27b85a0ca08bea0e
|
|
0ed8dbf042
Change-Id: Ibfdc4e41675638d0fe1ff81e67286064829c7e3a
|
|
Change-Id: Iddb8478deffb744e5d0d586ab5174927a3a667c3
|
|
Change-Id: Icd749280d34eaf1deb19d7f73a0ee23db7828706
|
|
0ed8dbf042
Change-Id: Id4458dba85b590cc2aed4dedf1ccfd3f93f43f29
|
|
Change-Id: I694e87918f034d686a969e0d61a2a3e5f7cb2b9c
|
|
Boundary check for NLMSG length is not present, which would
result in OOB read of the buffer during nla parse. Hence add
required boundary checks before processing ctrl msg from
wifihal clients. Also allocate required buffer for control
events to the size of received nlmsg data
Bug: 149836664
Test: Manual - Basic wifi sanity test
CRs-Fixed: 2628103
Change-Id: Idc971a343f30e41f359180757f108a8c0dfe2a7e
Signed-off-by: Vinay Gannevaram <quic_vganneva@quicinc.com>
|
|
This changes is to avoid the buffer-overflow in register_monitor_sock()
where "nreg->monsock" is getting filled with the data of length
"ctrl_msg->monsock_len" without the validation of length. Address this
issue by have a length check done before filling the buffer.
Bug: 149836664
Test: Manual - Basic wifi sanity test
CRs-Fixed: 2604404
Change-Id: I36b7ac274bd4f92eceabd5bd6534c73ae5a9ae73
Signed-off-by: Vinay Gannevaram <quic_vganneva@quicinc.com>
|
|
In send_nl_data() function, the size of ctrl_msg can
be greater than size of nl_msg structure. This can
cause buffer overload due to out-of bound write in
nl_msg->nm_nlh. Added a check for length of ctrl_msg
to avoid the out-of-bound write.
Bug: 149836664
Test: Manual - Basic wifi sanity test
CRs-Fixed: 2605058
Change-Id: I73032dac6ce2f2e9ee7ede18b45b11a2b3f92053
Signed-off-by: Vinay Gannevaram <quic_vganneva@quicinc.com>
|
|
9ea329ed12
Change-Id: I121f7aa15028164dbe94773f362f1f40b3e34b3b
|
|
All the users of this header are already using it as if
LOCAL_EXPORT_C_INCLUDE_DIRS has been set.
Bug: 130719878
Test: m libwifi-hal-qcom
Test: treehugger
Change-Id: I3c335b016208bfa5064f4a02f9f16447c6d29351
|
|
This commit replaces "0x%lx" with PRIx64 macro for better portability of
handling 64bit values in log messages.
Bug: 144576287
Test: Builds passes successfully
Change-Id: I8b8cd56c874f87e477b7a78b2c0e5197d77df483
|
|
This commit makes some adjustments to accomodate the change of the
feature_set from being an int (32 bits) into becoming a long (64 bits)
Bug: 139354972
Test: Manual: Build succeeds
Change-Id: I665b57707fa5b22dd847023c02c92e8513c9caf6
|
|
Currently after processing of back to back Nan Event i.e.
QCA_NL80211_VENDOR_SUBCMD_NAN and QCA_NL80211_VENDOR_SUBCMD_NDP,
mNanVendorEvent is not clear properly. Due to this stale entry
of mNanVendorEvent been accessing.
Clear stale entry of mNanVendorEvent on completion of processing of Nan
Event.
Fixes: 141010724
Bug: 141010724
Test: CtsNetTestCases with hwasan
CRs-Fixed: 2537045
Change-Id: I1342924d8da6f275ca0c1b4e043470a4f0680ac4
|
|
into qt-r1-dev
am: f2edcbeb9b
Change-Id: I20557bfdbd095f1e1a8b47c8c860092cb7ad99c4
|
|
into qt-r1-dev
|
|
am: 320302d001
Change-Id: I0197c57263b03bfdfcbe20bd094ed68d4d58700b
|
|
am: 1b30bee315
Change-Id: Ieba7e9b42fa9e5695392263e3e096642bde01e5b
|