| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Bug: 2538623
Change-Id: I779044173a2fa0e9b851ea92aeec5ca25eb9c86f
|
|
|
|
|
|
|
|
|
| |
- Remove check for supplicant command: setScanResultHandling
- Add check for stopDriver
- Fix condition for rest of commands
Bug: 2771298
Change-Id: I1262b2228814c4f6a4c251390af2f445c5f47234
|
|\
| |
| |
| |
| |
| |
| | |
Merge commit '9f2a0115220c8e288c32d8cf90e235a2f2be115b' into kraken
* commit '9f2a0115220c8e288c32d8cf90e235a2f2be115b':
Remove indefinite DHCP retries
|
| |
| |
| |
| |
| |
| |
| |
| | |
Replace indefinite DHCP retries
with a sufficiently large retry count
Bug: 2761057
Change-Id: I4570b74476aa78c71c73e06c2aafac33d924e867
|
|/
|
|
|
|
|
| |
This is to support multiple headsets at the same time, atleast
at the framework level.
Change-Id: I91d05c6c2828c9a09d00806d5e79f1e9c9c7cf84
|
|
|
|
|
| |
Bug: 2600111
Change-Id: Ifaa63cd7c05dfa27fbd851de6f7e92f7fc38907b
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Due to message removal, wakelock could be held forever.
Do a timer only based wakelock release until we do this
more cleanly in ConnectivityService for later release.
Also, add an optimization to prevent use of wakelocks when driver is
already stopped.
Bug: 2529883
Change-Id: Ia1c2ddd44213ef3aa609855613bf155945bef8e4
|
|
|
|
|
|
|
|
| |
mWifiState was split up to speed up access
of wifi state when things are synchronized.
Use a single atomic state variable throughout.
Change-Id: I345981e3ffa3d2b0b7e70a5c6efc43185cba9062
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previous it was synchronized on the same lock that was held while
transitioning wifi states, effectively making it impossible to ever
find out about such states, and worse: making calls during transitions
block.
This manifested itself in a really laggy power control widget, which
I'm working on fixing in general in a separate change.
Change-Id: Ia34af98c73d7f5662e41360a213df87f6adaf130
|
|
|
|
|
| |
Bug: 2525763
Change-Id: If2f334187af26e0352d754d7e00a9ca3efc5dffa
|
|
|
|
|
|
|
|
|
| |
Remove AP persist settings
Add new netd interface
Handle errors
Handle AP config change
Bug: 2413908
Change-Id: I31a1221ef5479da8d4a2620f0f0ee0b62539bc69
|
|
|
|
|
|
|
|
| |
Driver commands should be issued when driver has started.
Supplicant commands should be issued when Wi-Fi is enabled
Bug: 2339709
Bug: 2371609
Change-Id: I9ba6ddfa0cf4c4b8ca049b0eb7eaaa8edb42bad1
|
|
|
|
|
| |
Bug: 2339709
Change-Id: Ibb2eeb09b83dc24897116ea15a9016f81a08d442
|
|
|
|
|
| |
Bug: 2421638
Change-Id: Ic5ea8f7560a7fe5e1b0769daa5d92cc33eefc692
|
|
|
|
|
| |
Bug: 2274930
Change-Id: Ia1eafe422f6e18b826e53e72a306dd2961b948a6
|
|
|
|
|
|
|
|
|
|
| |
The IP state was not being refreshed when the supplicant transitions
from COMPLETED to ASSOCIATED to COMPLETED. This can lead to
a connected state with no real connection due to old IP settings.
The fix refreshes IP on each connection.
Bug: 2329261
Change-Id: I38cd56369ee2d8ab3e0f06f5c9f5712b9b2f35a0
|
|
|
|
|
| |
Bug: 2399119
Change-Id: I481cfbfb11fb453659134004aafa3df236ba050b
|
|
|
|
|
|
|
|
| |
We currently disable networks upon too many reconnects. This leads to asking
the user input for reconnects. Blacklist it instead.
Bug: 2129037
Change-Id: I23d69daf3964c066ed7f70d32fefb81016f19aa2
|
|
|
|
|
|
|
| |
Some of the native calls were left unsynchronized in the framework. Pre-empted IOCTL call
interrupted by another call from the framework cannot be handled in the driver.
Bug: 2310455
|
|
|
|
|
|
| |
Use this in various places where it should serve no purpose to deliver
both broadcasts. This is intended to reduce somewhat the flurry of
broadcasts that we churn through during boot.
|
|\
| |
| |
| |
| |
| |
| | |
Merge commit '68b15b68497e373fe63a13f11e18349a738a7989' into eclair-mr2
* commit '68b15b68497e373fe63a13f11e18349a738a7989':
Porting change I064698b3 from mr2 to eclair.
|
| |
| |
| |
| |
| |
| | |
Add support for "HANGED" driver state - we'll reboot the driver when requested (by the driver).
bug: 2310372
|
| |
| |
| |
| |
| |
| |
| |
| | |
Waits 15sec after getting a DISCONNECTED or INACTIVE state. Also tracks
the number of supplicant state transitions. After the period is up
if we haven't received any more supplicant state transitions, it requests a scan.
bug: 2168081
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The driver can now report to us that they are hosed and we'll shut
down wifi and restart it - only to be used as a last resort.
Also fixing synch problem with updateWifiState.
bug: 2173119
|
| |
| |
| |
| |
| |
| | |
Gets rid of the old method which generated Log noise.
bug: 1973855
|
|/
|
|
|
|
|
|
| |
Waits 15sec after getting a DISCONNECTED or INACTIVE state. Also tracks
the number of supplicant state transitions. After the period is up
if we haven't received any more supplicant state transitions, it requests a scan.
bug: 2168081
|
|
|
|
|
|
|
| |
The driver will bring up the interface when needed and bring it up
at reset time has caused some problems.
bug: 2144870
|
|
|
|
| |
The field identity should not be "anonymous_identity".
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Problem:
Connect to AP with WEP security.
Change AP's WEP password.
Phone will quickly reassociated (less than in 5 sec) and will not restart full "reconnect".
Indication will show "connected" but it will be no network access.
Fix: (for non-static IP only)
With WEP mode, the association could be successful if we change the AP's WEP password. However,
the phone wont be able to send/recv any traffic with the old password. What we could quickly
fix here is to send the DHCP request if the network is not configured to use static IP. We
can do nothing but ping access point to see if it we are able to communicate with it.
|
|
|
|
|
|
|
|
|
| |
Only do the full signal strength calc when we need to, otherwise use
an approximate value.
This change was introduced a while back but was lost in some merge.
bug: 2125543
|
| |
|
|
|
|
| |
+ refactor the enterprise fields.
|
|
|
|
|
|
|
|
|
| |
The supplicant can take up to 15 seconds to start - setting the number of wifi channels
immediately after requested wifi start often will fail.
Changed to set the number of channels when the supplicant is reported as alive.
bug:2083601
|
|
|
|
| |
Change-Id: I2205eebae419eaf4a0992c9f5b7cd807eb843fe1
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Split BluetoothDevice into BluetoothDevice and BluetoothAdapter.
BluetoothAdapter: Represents the local BT adapter. Operations on the local
adapter (start a scan, etc).
BluetoothDevice: Represents a remote BT device. Operations on remote devices
(pair, connect, etc).
IBluetoothDevice.aidl -> Bluetooth.aidl
BluetoothDeviceService.java -> BluetoothDeviceService.java
TODO:
Javadoc
|
|\
| |
| |
| |
| |
| |
| | |
Merge commit '41c104339951e0e5e78240e1f48455c21b3ba5fd'
* commit '41c104339951e0e5e78240e1f48455c21b3ba5fd':
Fix MulticastLock API to match WifiLock API.
|
| |
| |
| |
| | |
Adds option to make lock refcounted or not. Fixes 2017680.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
driver.
This fixes a problem that resulted in the interface not being up when resuming
an authenticated connection.
Signed-off-by: Mike Lockwood <lockwood@android.com>
|
|\ \
| |/
| |
| |
| |
| |
| | |
Merge commit '356d4a14aa96cf52c16f7a4d381044ce28b01af3'
* commit '356d4a14aa96cf52c16f7a4d381044ce28b01af3':
Add the phase2 field for EAP WiFi configuration.
|
| | |
|
|\ \
| |/ |
|
| |
| |
| |
| | |
Signed-off-by: Mike Lockwood <lockwood@android.com>
|
| |
| |
| |
| |
| |
| |
| | |
We now only bring it down briefly and bring it back up again when
we want to reset the interface.
Signed-off-by: Mike Lockwood <lockwood@android.com>
|
| |
| |
| |
| |
| |
| |
| | |
Apparently this code was not 100% correct, and now the Wifi driver is responsible
for setting the scan mode to active for hidden networks, so this is no longer needed.
Signed-off-by: Mike Lockwood <lockwood@android.com>
|
|\ \
| |/
| |
| |
| |
| |
| | |
Merge commit '463aacfbd220c9ef2f6ed915d54092289fdac05b'
* commit '463aacfbd220c9ef2f6ed915d54092289fdac05b':
wifi: WifiManager.startScan() will now do passive scans by default.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Active scans will only happen if a hidden AP is in use, or if the new method
WifiManager.startScanActive() is called.
This fixes some audio playback problems with bluetooth A2DP.
Signed-off-by: Mike Lockwood <lockwood@android.com>
|
|\ \
| |/
| |
| |
| |
| |
| | |
Merge commit '699ca3f2518360ea3250ff5a0e5d39e122c64a91'
* commit '699ca3f2518360ea3250ff5a0e5d39e122c64a91':
Add password field for WiFi configuration.
|
| |
| |
| |
| |
| |
| |
| | |
1. the certtool.h is modified for avoiding the side effect,
for saving the configuration with wpa_supplicant.
2. put the loadLibrary back in CertTool.java
3. Fix incorrect JNI declarations.
|
|\ \
| |/
| |
| |
| |
| |
| | |
Merge commit '3bc64a2cb97da7bf7dc203416ad6c7d86c5925fc'
* commit '3bc64a2cb97da7bf7dc203416ad6c7d86c5925fc':
WifiManager: Limit the number of WifiLocks that can be active simultaneously.
|
| |
| |
| |
| |
| |
| |
| | |
This is to prevent a bad app from using up all the global references in the
system process due to allocating too many DeathRecipients.
Signed-off-by: Mike Lockwood <lockwood@android.com>
|