| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
| |
Code search says these are the only two files that use it. The
tracker will be resurrected in a slightly different form in
frameworks/opt/net/ethernet.
Bug: 14993642
Bug: 14981801
Change-Id: I2477668ca78dfe46661dda1d97c7f786fd7eba35
|
| |
|
|
|
|
|
| |
Also remove unused LinkSocket and LinkSocketNotifier.
bug:13885501
Change-Id: Id426e31b201fa4f29109b5fea485d8efb34519d3
|
| |
|
|
| |
Change-Id: If2a5c8bda2cb24c044903df64705bf7b3587fee1
|
| |
|
|
| |
Change-Id: I154f151ca676e1e4137eb8210fd377659cfc7867
|
| |
|
|
| |
Change-Id: Ib01333894cc4dfbe223ef26e6eb815812d32092d
|
| |\
| |
| |
| |
| |
| |
| | |
be tuned seperately issue"
* commit '15b3efc9620662d83b66d75c663bb893ca0bc6e9':
fix ethernet and wifi can't be tuned seperately issue
|
| | |\
| | |
| | |
| | |
| | |
| | |
| | | |
seperately issue"
* commit 'ecfae4f899873f224e1aeed076dc8a41f8884487':
fix ethernet and wifi can't be tuned seperately issue
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
ethernet especially Gigabit enet use the same tune data with wifi
won't reach best performance.
Change-Id: Iac50ba47904c76ff7b5b8ad226e83451359ec534
Signed-off-by: Jianzheng Zhou <jianzheng.zhou@freescale.com>
|
| |/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Note that this CL does not change any behaviour.
At the center of this change is
CaptivePortalTracker#detectCaptivePortal(), which does nothing
except call back into ConnectivityService. Removing it allows us to
simplify code in ConnectivityService. It also allows us to remove
ConnectivityService#captivePortalCheckComplete which was only ever
called in response to this method.
While this does not change any behaviour, it preserves existing
bad behaviour, i.e, that the CAPTIVE_PORTAL_CHECK NetworkInfo
state does not correspond to actual captive portal detection.
We transition into that state and immediately (and unconditionally)
out of it and into CONNECTED.
Change-Id: Ib3797f956d2db5e3cacaaa53e899d81aa8e958af
|
| |\ \ |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This change starts tracking traffic quality data for WiFi and mobile
networks. The quality is tracked based on incidental traffic, and not
on specific measurements. Theoretical bandwidths are hard-coded, as
well as sampling interval; although sampling interval can be changed
by setting a system policy.
Bugs filed to remove shortcomings of this change -
10342372 Change LinkInfo name to something better
10342318 Move hardcoded values of MobileLinkInfo to resources
so they can be updated without changing code
Bug: 10006249
Change-Id: I83d8c7594da20fe53abbd5e1f909b1f606b035bb
|
| |/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently BaseNetworkObserver is in frameworks/base/services, but
there is code in frameworks/base/core that could use it. This
code typically extends INetworkManagementEventObserver.stub
because BaseNetworkObserver is not available.
Move BaseNetworkObserver to frameworks/base/core without changing
its package name, and use it to simplify two callers. The third
caller, Tethering, is much larger, and I'm not sure it's
appropriate to change it.
Bug: 10232006
Change-Id: Ifc0f2e619e3424e27e35730c048a1cc523df345e
|
| |\ \
| |/
|/|
| |
| |
| |
| | |
enable fail fast." into jb-mr2-dev
* commit '02eab434ad9faa25291e63023b7e66698ca457a8':
If in a mobile captive portal is detected enable fail fast.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When captive portal checking completes pass back the result.
This is used to enable/disable failing fast for mobile. When
failing fast is enabled we don't check for data stalls and thus
won't be continually trying to do recovery operations, such as
restarting the radio.
Bug: 9462512
Change-Id: I0dea0eee519f8ee7f94e79d40e82c18f30d7fe2e
|
| |\ \
| |/
|/|
| |
| | |
* commit 'a34b531bb6d2f2923f474b8c0a0c09149303e5d3':
EthernetDataTracker: Set network available on connected
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This change fixes state mismatch found in the bugs below,
and makes the code look consistent with disconnect().
In startMonitoring(), reconnect() is called before mNMService.registerObserver().
Then dhcpThread.run() could finish without getting interfaceLinkStateChanged() callback.
In such a case, mNetworkInfo has connected but not-available state - which is contradictory.
Bug: 8113933, 8291000, 8291039
Change-Id: I29d74f8d09dcffb97d6c9f8198811bca046fef0c
|
| | |
| |
| |
| |
| |
| |
| |
| | |
Allows the external BT stack the means to communicate with
ConnectivityService during reverse tethering.
bug:8445208
Change-Id: Ice7dfb0b50c9481d359aed14a51372878185171c
|
| | |
| |
| |
| |
| | |
Bug: 8276725
Change-Id: I55a107fea78f9cc84f15bd6c75292138fb35a1b6
|
| |\|
| |
| |
| |
| | |
* commit 'fd0b25875c3fe005329d412a2500da1951ff6a7b':
EthernetDataTracker: Start DHCP on Link Up
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The behavior of mLinkUp has previously changed so that it only
represents the link state. This changes fixes
interfaceLinkStateChanged() so that it triggers a reconnect based
only on the present link state.
Change-Id: I950e04e1f5b5019d59d3b9530f369f8b8f13134a
Signed-off-by: Adam Hampson <ahampson@google.com>
|
| |/
|
|
|
|
|
| |
First step in accepting a set of patches.
bug:6799630
Change-Id: I6c894c60aeb3022960c2aaa45451bb1dde2b493b
|
| |\
| |
| |
| |
| |
| |
| | |
link up based on flags"
* commit '5d33e76cff65245d03f592a3dbcbf2e17a2f1581':
EthernetDataTracker: indicate link up based on flags
|
| | |
| |
| |
| |
| |
| |
| |
| | |
Indicate link up state based on flags/interface up, and not on IP address.
This is for ethernet interfaces that already exists.
Change-Id: Ib342d519c483bbb2dfa08cfac2c0c1a288cee7c0
Signed-off-by: Vishal Mahaveer <vishalm@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We now notify the user of a captive portal before switching to the network as default.
This allows background applications to continue to work until the user confirms he
wants to sign in to the captive portal.
Also, moved out captive portal handling out of wifi as a seperate component.
Change-Id: I7c7507481967e33a1afad0b4961688bd192f0d31
|
| | |
| |
| |
| |
| |
| |
| |
| | |
The activity notification is received from netd, an intent
DATA_ACTIVITY_CHANGE is then raised for other part of the system to
consume.
Change-Id: Idfcc4763c51c5b314c57f546c12557082f06bebf
|
| |\ \
| |/
|/|
| |
| |
| |
| | |
interface
* commit '1d28fef92e516b0144d7844413194ba9e953b317':
Stop any running DHCP client before bringup up an Ethernet interface
|
| | |
| |
| |
| |
| |
| |
| |
| | |
If the system server is restarted, the dhcpcd_eth0 service may have already
been started by the previous system server, and the new system server's
attempts to start the dhcpcd will do nothing.
Change-Id: Icfd480a5369eb9c7dfe86b43190294871a915529
|
| | |
| |
| |
| |
| |
| |
| |
| | |
Previously we were starting DHCP as soon as the interface was added,
which is at boot on devices with builtin ethernet.
Change-Id: I726c2b139cbf91455d159d7cbcd08e24a336942b
Signed-off-by: Mike Lockwood <lockwood@google.com>
|
| | |
| |
| |
| | |
Change-Id: Ic645396e81d1481d84061945eb9721f0eeb151eb
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When a cable was unplugged, we were telling the driver to release the
ip address so if a cable on a different network was plugged in, it would
still try to use it's old ip address on the new network, which probably
didn't work.
Also, we didn't notify ConnectivityService about the state change in
the unplug case. Some of this was done in the interface removed case,
but we never remove the interface in Tungsten, just unplug. So refactor
the common disconnect code into a disconnect() function that's called
by both the link status change (unplug) and interface removal (only applies
to things like USB ethernet dongles) cases.
Change-Id: I19d7ceb540d6a0e24df1b55314a13af78f1e593d
Signed-off-by: Mike J. Chen <mjchen@google.com>
|
| | |
| |
| |
| |
| |
| |
| |
| | |
We don't get link status info if it's down. If the device has been configured
for ethernet and we see the interface appear, we should make it active.
bug:6444395
Change-Id: Ibc233ab12942b069d7db9c0671936798c5161659
|
| | |
| |
| |
| |
| |
| |
| | |
Previously we were starting DHCP as soon as the interface was added,
which is at boot on devices with builtin ethernet.
Signed-off-by: Mike Lockwood <lockwood@google.com>
|
| | | |
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When a cable was unplugged, we were telling the driver to release the
ip address so if a cable on a different network was plugged in, it would
still try to use it's old ip address on the new network, which probably
didn't work.
Also, we didn't notify ConnectivityService about the state change in
the unplug case. Some of this was done in the interface removed case,
but we never remove the interface in Tungsten, just unplug. So refactor
the common disconnect code into a disconnect() function that's called
by both the link status change (unplug) and interface removal (only applies
to things like USB ethernet dongles) cases.
Signed-off-by: Mike J. Chen <mjchen@google.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add policy controls to NetworkStateTracker which are combined with
other user preference and internal flags to decide if data connection
should be established. Better locking around enabled flags.
When data network would be over limit, proactively disable data on
that network. Enable when policy is snoozed or when cycle resets.
Track and dismiss notifications from now-stale policies.
Bug: 4587023, 5178147
Change-Id: Ibfcc9f73cda7c369209af701b46eddd3d1943f2d
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a 1st pass at receiving events that indicate
some quota has been reached e.g. warning quota, data collection quota,
cutoff quota,...
It needs:
- new kernel with quota2 logging support
- new net:bandwidthcontroller that supports
. quota2,
. setting alerts.
- new NetlinkEvent/NetlinkManager/NetlinkHandler to process
NETLINK NFLOG messages.
Change-Id: Ibfbb13512c5350cdee0e544ec14caa6f59812409
|
| |
|
|
|
|
|
|
|
|
|
|
| |
After unreverting the linkstate change patch, hook up notification handlers
that didn't exist when the first patch was created, like
EthernetDataTracker.java and Vpn.java.
For the observers that handle interfaceStatusChanged(), I made
interfaceLinkStatusChanged() call it so they both do the same thing.
Change-Id: I0077e5e5f48f3932ba98f5bf363243892f2de6cc
Signed-off-by: Mike J. Chen <mjchen@google.com>
|
| |
|
|
|
|
|
|
|
|
| |
Devices with always present wired ethernet interfaces (as opposed
to usb ethernet dongles) don't have the interface added and removed.
Instead, it is present already at boot time. Change EthernetDataTracker
to support the already existing at boot time case.
Change-Id: I7b8e938f6b07aabd14f00ace55b35b7a05ea5a38
Signed-off-by: Mike J. Chen <mjchen@google.com>
|
| |
|
|
| |
Change-Id: I9ac1c28de4f3f11f84b04923d4547ab64dd57f23
|
|
|
Change-Id: Ie75cbcdb1b0c8965b59b89675a05175a1e9ebdd6
|