Age | Commit message (Collapse) | Author |
|
15873656, 15873657, 15873637, 15873638, 15873639, 15873339, 15873340, 15873341, 15873342, 15873343, 15873344, 15873345, 15873346, 15873947, 15873948, 15873949, 15873950, 15873951, 15873952, 15873953, 15873954, 15873955, 15873956, 15873957, 15873958, 15873959, 15873960, 15873961, 15873642, 15873122, 15873323, 15873123, 15873324, 15872698, 15872700, 15873987, 15873962, 15873963, 15873964, 15873965, 15873557, 15873558, 15873816, 15873643, 15873817, 15873660, 15873559, 15873988, 15873644, 15873645, 15873646, 15873325, 15873661, 15873662, 15873663, 15874027, 15874028, 15874029, 15874030, 15874031, 15874032, 15874033, 15874034, 15874035, 15874036, 15874037, 15874038, 15874039, 15874040, 15874041, 15874042, 15873560, 15873561, 15873854, 15873855, 15873562, 15873563, 15873564, 15873565, 15873566, 15874047, 15874048, 15873326, 15873124, 15874043, 15872622, 15872623, 15873857, 15872701, 15872702] into sc-d1-release
Change-Id: I3b67239b99a1795980fbe1dacce2899f77aff01c
|
|
A while loop and condition check for the value of a type to be 0
when in fact since the value.len is arbitrary it could make the
remaining length "less than 0" and since the type is unsigned it'll
never be "less than 0."
Use signed type for loop and conditional checking.
Additionally, make sure the value.len when used to read an array is not
more than the remaining length of the data.
Bug: 197536150
Test: poc application
Tag: #security
Change-Id: I20d66ddd1055577d7d39aba447233c19081bb789
(cherry picked from commit 1da56d1c815aa4854aa42f721732070333e5e924)
|
|
Change-Id: If14606a8e064c8cc2d36d5270c785c23c18c22a1
|
|
|
|
Change-Id: I7997b286e7272226365bdeafe762d0124faca5df
|
|
Bug: 175425864
Tag: #compatibility
Test: Switch A2DP active device manually
Ignore-AOSP-First: cherry-pick from aosp to upstream branch
Change-Id: Icf29d8e5c557c927bffe9d49188bd0cf98b07120
|
|
Inquiry is not stopped cause of clearing of BTM_BLE_INQUIRY_MASK.
This issue came from the follwing patch.
https://android-review.googlesource.com/c/platform/system/bt/+/1402183
Bug: 195908804
Change-Id: I3361c924c9445d6aae2856f41b732fca22951f76
Tag: #refactor
Ignore-AOSP-First: cherry-pick from aosp to upstream branch
Test: compile & verify basic functions working
|
|
Change-Id: I9790c61aa4aa4ca5a661759599c65dbcedb8a703
|
|
|
|
When sniff is enabled, KDDI car kit may disconnect the ACL connection when SCO connection disconnect.
Tag: #compatibility
Bug: 194029299
Test: 1) Connect with KDDI car kit
2) Check Write Link Policy Command, enable sniff mode should be false
Change-Id: I597684cc2dc709167a789e2a1de3866bf9aab872
|
|
Change-Id: Ie56a9db7594289057fb748d9a15edd7eb822d45b
|
|
To avoid ACL link go into sniff mode during a2dp streaming by remote
request, disable sniff mode by link policy when AVDTP start, and
enable when AVDTP suspend, stop or disconnect.
Tag: #compatibility
Bug: 193206591
Test: audio resume after MO/MT call
Change-Id: I0af3223e0722b0343e8d091985e3e23606104268
Merged-In: I0af3223e0722b0343e8d091985e3e23606104268
|
|
Change-Id: I6af93e85a83a7855d0dd05a10a128392fd93e1e6
|
|
|
|
sc-dev
|
|
Change-Id: I62d8d9eb2ce09bcc16502e9f01c1847dc0107934
|
|
Bug: 194432570
Test: Manual test app; nRF connect
Tag: #feature
Change-Id: I17a1a8b1ba2b305527ae74bb8ad5f6e1258e870e
Merged-In: I17a1a8b1ba2b305527ae74bb8ad5f6e1258e870e
|
|
Bug: 192154257
Test: gd/cert/run
Tag: #refactor
Ignore-AOSP-First: Possible CP candidate
Change-Id: Ia916e427cb00e029ecb13a56255219fb2aba8500
|
|
We should register security record when we are server, so the condition
should be the same as "vpsm == psm", as in line 412 below. So we should
use "pL2CA_ConnectInd_Cb != nullptr" (server has to have an incoming
connection handler), OR "psm < LE_DYNAMIC_PSM_START" (Fixed PSM service,
as it's treated like this for vpsm).
Tag: #stability
Bug: 193142224
Test: CtsVerifier LE COC client and server
Change-Id: I627e4dcd0aca4c113966952b53fad2be0f9d7104
|
|
|
|
- When we are client, we don't need to register security during L2cap
registration.
- When we are done, we clean up security record by PSM; Service ID might
be invalid.
- Once the btif topshim facade initial commit is done, we can repro this
and verify the fix with test automation.
Tag: #stability
Bug: 193142224
Test: CtsVerifier LE COC Client for many times; later we will have an
automated test for this
Change-Id: I90fd23ce26c65ca3314e0754a2630d3f63c5d5d8
Merged-In: I90fd23ce26c65ca3314e0754a2630d3f63c5d5d8
|
|
Change-Id: I0fd7ca9309b5669c67f71766f2aa871610975faa
|
|
When we receive a non-SUCCESS command status for pm change, we should
not stop here; there might be some other pending pm changes in the
queue. We need to try the remaining ones.
Test: Disconnect a device when there is a pending PM change
Tag: #stability
Bug: 180842489
Bug: 184095368
Change-Id: I31a6f96cd3b47e671141c29801f4d7397787fc51
Merged-In: I31a6f96cd3b47e671141c29801f4d7397787fc51
|
|
Change-Id: Ic68e813b824c613d759d494b47ca068bb0f799a8
|
|
Original change: https://android-review.googlesource.com/c/platform/system/bt/+/1740854
Bug: 191431228
Tag: #refactor
Test: gd/cert/run --device
Ignore-AOSP-First: Cherry-pick to release branch
(cherry picked from commit b9eb2c51c7a7af64da6f622efbc2f822c3c000c8)
Merged-In: I25e3a8aa16dce375fed79133b1e1561d914c6a8d
Change-Id: I25e3a8aa16dce375fed79133b1e1561d914c6a8d
|
|
Change-Id: If1a890f3296a9f6989de8e5729585d0850f6eb35
|
|
With this patch we make sure that eatt_dev is created always when EATT
is supported on the peer side, no matter what is the connection role.
It will allow to accept ecoc channels when Android device is peripheral
device
Bug: 159786353
Bug: 191313013
Tag: #feature
Test: atest --host net_test_eatt
Sponsor: jpawlowski@
Merged-In: I4d735bc4a2d74f637e9c7f7819e10659af9b0fbb
Change-Id: I4d735bc4a2d74f637e9c7f7819e10659af9b0fbb
|
|
It fixes:
signal 6 (SIGABRT), code -1 (SI_QUEUE), fault addr --------
Abort message: '[FATAL:eatt_impl.h(142)] Check failed: eatt_dev->eatt_tcb_.
Crash could happen when Android is connects at first as a central device
connection to peer which supports EATT and after reconnection it is peer
initiating ecoc (e.g. Android is peripheral or peer is faster with ecoc
connection request). At that point eatt_dev is there but eatt_tcb_ is
not.
Bug: 159786353
Bug: 191313013
Tag: #feature
Test: atest --host net_test_eatt
Sponsor: jpawlowski@
Merged-In: Ic5e8c027dc45eec942457f09c987ec64cfafb761
Change-Id: Ic5e8c027dc45eec942457f09c987ec64cfafb761
|
|
Tag: #feature
Bug: 159786353
Sponsor: jpawlowski@
Test: compile & manual testing
Bug: 191313013
Merged-In: Ic16d962949e3aa51ed1ab5ce06cf9aa1acebc726
Change-Id: Ic16d962949e3aa51ed1ab5ce06cf9aa1acebc726
|
|
This is needed for storing Enhanced ATT support.
Tag: #feature
Bug: 159786353
Sponsor: jpawlowski@
Test: compile & manual testing
Bug: 191313013
Merged-In: Ic37b668b91ab6c830d780f70db703a5d9be11677
Change-Id: Ic37b668b91ab6c830d780f70db703a5d9be11677
|
|
Bug: 159786353
Tag: #feature
Test: atest --host net_test_eatt
Sponsor: jpawlowski@
Bug: 191313013
Merged-In: Ie4872f1c51445df59f2f68b127454d5f4d4e9ad0
Change-Id: Ie4872f1c51445df59f2f68b127454d5f4d4e9ad0
|
|
Bug: 159786353
Tag: #feature
Test: compile
Sponsor: jpawlowski@
Bug: 191313013
Merged-In: Ia07b894a37a648d2a8aec0539e6fb75000a2c999
Change-Id: Ia07b894a37a648d2a8aec0539e6fb75000a2c999
|
|
Change-Id: Ic319a14d79ad357ae680d414ce3257e4772da9bc
|
|
|
|
Change-Id: I7bca1a1865a806f62d8b04b040406adcaf3ebc7c
|
|
Tag: #feature
Bug: 187165224
Test: Manual
Merged-In: I260c967de0f4656ee852a098a98c9ceb0e6dfbde
Change-Id: I260c967de0f4656ee852a098a98c9ceb0e6dfbde
|
|
Change-Id: I17bfcc834241660d34b9bbf113f417ce02d98b67
|
|
Bug: 192475074
Test: Manual, test app.
Tag: #feature
Change-Id: I5086cbe6c0d2ab3346a956e22df36725a9e93530
Merged-In: I5086cbe6c0d2ab3346a956e22df36725a9e93530
|
|
sc-dev
|
|
Change-Id: I5299f6d13823a8a7612fc47a27d3dd55c7d0d1e3
|
|
* changes:
OOB: Replace 'memcpy' with assignments
OOB: Cancel advertiser and replace
OOB: Start advertising when OOB data is generated.
OOB: Connector; Set name of remote from OOB data
OOB: Store keys and data generated for local OOB use.
OOB: Utilize SMP state machine to generate the OOB data.
OOB: Rework create bond out of band flow to not use btif_dm_cb_create_bond
|
|
Fixes a lint warning
Bug: 178007935
Test: Compiles
Tag: #feature
Change-Id: Id5e0d41085f9415daadb261a01f99200cb8f90e0
Merged-In: Id5e0d41085f9415daadb261a01f99200cb8f90e0
|
|
Bug: 192298941
Test: Manual, test app and nRF connect.
Tag: #feature
Change-Id: Ia9a538fea6ac58e1487523447b22396c0caead8b
Merged-In: Ia9a538fea6ac58e1487523447b22396c0caead8b
|
|
This affords the stack the capability to provide the proper
random address and address type to the user via the OobDataCallback.
Previous to this commit, the data defaults to the public address.
This advertisement will time out, and will clear OOB data upon timeout.
Bug: 192084169
Test: Manual, test app.
Test: m bt_host_test_bta
Test: m net_test_btif_stack
Tag: #feature
Change-Id: Ia24c10e7209027b8b070b9196caba7e394ee6dcb
Merged-In: Ia24c10e7209027b8b070b9196caba7e394ee6dcb
|
|
There are 4 total "devices" listed in Settings.
Each shows a mac address, if a name doesn't exist it won't
show in things like the share a file dialog.
This names 1/4 of those addresses. The LE Device (Advertiser)
displayed on the Connector's Settings->Connected devices list.
Bug: 192093473
Test: Manual, test app
Tag: #feature
Change-Id: I94024413e962f4742f6fa71c495d042aaede5ffb
Merged-In: I94024413e962f4742f6fa71c495d042aaede5ffb
|
|
Bug: 190395922
Test: Manual, test app
Tag: #feature
Change-Id: Ib88abb866ce08b99c7fbf17ac9cfb85ee0f3441e
Merged-In: Ib88abb866ce08b99c7fbf17ac9cfb85ee0f3441e
|
|
Finish the SMP implemention for OOB in the legacy stack and
connect callback pipeline to return data to the caller.
Bug: 178007935
Test: Manual; test app
Tag: #feature
Change-Id: Id4fe6b29a1db3ec50d37b042ea67cebfa9eba4e1
Merged-In: Id4fe6b29a1db3ec50d37b042ea67cebfa9eba4e1
|
|
Instead we just call BTA_DmBond directly since this is an OOB pairing scenario
Bug: 178007935
Test: Compiles, Manually with test app
Tag: #feature
Change-Id: Ie9d7cf20c18b2ccac322dabad141135e8a18bf93
Merged-In: Ie9d7cf20c18b2ccac322dabad141135e8a18bf93
|
|
As it stands, both newavrcp target and avrcp controller devices are
configured as not "passive." This is an Android BT stack convention that
AVCTP uses as a policy for handling unlikely connection collisions. When
passive, a device will allow an incoming connection to take the place of
it's outgoing connection, assuming the connection. Non-passive devices
are strict in only allowing their outgoing connection to stay up.
When two non-passive devices connect with each other, there's a 10-15%
chance that the timing of events will cause each device to disconnect
each other's connection, resulting in no connection being set up.
AVCTP Section 5.1 says the target should allow the controller's
connection to succeed. Configuring newavrcp target to be passive puts us
more in line with that requirement.
Note this is a temporary work-around for a larger effort around the
AVCTP connection collision handling that will probably need to take
place.
Tag: #stability
Bug: 179292409
Test: atest BluetoothInstrumentationTests; -- Also connect phone to an
AVRCP controller device and wait for an unlikely connection collision.
Ensure the phone allows the CT to connect over its own connection.
Change-Id: I6861c78ba594d4e4ccfcf8ce9adec109f1b2d270
Merged-In: I6861c78ba594d4e4ccfcf8ce9adec109f1b2d270
|
|
Bug: 168712382
Tag: #security
Test: gd/cert/run
Ignore-AOSP-First: Security
Change-Id: Iae823e45675d46d8ca037157e516cc2f94fadfab
|