This reverts commit 55ccfe19e2.
Reason for revert: this must be introducing some sort of race
it appears to cause
atest EthernetTetheringTest 'NetdBinderTest#TetherForwardAddRemove'
to no longer reliably pass.
Change-Id: I5281ab3f42c5ce268d97a12db24a6768db3f4354
Maciej Żenczykowskiauthored and Gerrit Code Reviewcommited 2023-10-28 02:10:52 +00:00
1. Move it to header file so that it can be reused by others.
2. Correct the return type from int to bool.
3. Replace __always_inline by inline to avoid -Werror,-Wunused-function.
Bug: 288340533
Test: build
Change-Id: I9062686d9c2f98c2d24e4673f82b1732b180ffc4
This is no longer used out of this file, and should not be.
Test: Build
Change-Id: I531191b4afbda39a5fbaf1f483d13068f86d17f1
Merged-In: Ief0a79883bcc2c5493807c548cb71ef655abed23
Chalard Jeanauthored and Jean Chalardcommited 2023-10-27 06:53:36 +00:00
BPF needs upstream prefixes information to filter spoofing IPv6 source
addresses carried in downstream traffic.
We retrieve prefixes from upstream interface's LinkProperties and pass
it to the BpfCoordinator. Forwarding rules will also be updated when
upstream interface's IPv6 link addresses change.
Test: atest TetheringTests
Bug: 261923493
Change-Id: If8cfc3b191e520ca838654d1b5211ab9e9ec021d
As state transitions always occur after a message has been processed,
calling transitionTo only under processMessage makes the flow easier
to understand and less error-prone.
The risk of this change is very minimal since it is only about
handling the error, but tethering will stop regardless.
Test: atest TetheringTests
Change-Id: I56c6cf6cc989464ee84a8333ac131afc808a3d95
This can make it easier to adopt SM shim replacement in follow up change,
as both SyncSM and AsyncSM will process the same message without any
difference.
This change also fix a problem that bring up downstream before it's
IpServer completely disabled.
Previous:
CMD_NOTIFY_PREFIX_CONFLICT -> requestEnableTethering(false) ->
requestEnableTethering(true) -> IpServer stopped ->
reportTetherStateChanged(downstream inactive) -> IpServer started
-> reportTetherStateChanged(downstream active)
New:
CMD_NOTIFY_PREFIX_CONFLICT -> requestEnableTethering(false) ->
IpServer stopped -> reportTetherStateChanged(downstream inactive)
-> requestEnableTethering(true) -> IpServer started
-> reportTetherStateChanged(downstream active)
Test: atest MtsTetheringTestLatestSdk
Change-Id: Ibe22b300c56125359f4fa452bd01a0a2381fda23
When tethering restarts due to an IP conflict, it starts to
bring up downstream before its IpServer is completely disabled.
Test: atest TetheringTests
Change-Id: I0b0f0bf1bb9ac8421c05fe512d5082d095fa7180
proc file writes were only moved from .rc into bpfloader binary in U.
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I20eda71d4f4e16130401d1cfff239c7cd8ea9bf6
Maciej Żenczykowskicommited 2023-10-25 18:40:53 +00:00
(mostly to improve code coverage)
We can always revert if it ever turns out this is useful,
but I'd prefer for cgroup attach to be a once-at-boot
thing and handled either from bpfloader or the netd updatable
which are both directly C++ code.
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I79b5ca8a47388cb6b9189234942e74ab6056aab9
Maciej Żenczykowskicommited 2023-10-25 17:40:35 +00:00
Per the API review council, the classes should be "final" by default
Bug: 301076593
Test: build && atest CtsThreadNetworkTestCases
Change-Id: I85f0b4fabe5b1bcec2be052620b38a07c58181d7
The performPoll method in the NetworkStatsService can be called
for many reasons, such as RAT changed events, LinkProperties
changed events, open session method calls, forceUpdate method
calls, etc.
However, the current logging is not sufficient to trace down
the caller of performPoll. This makes it difficult to debug
frequent poll issues or performance problems.
Adding debug log to logcat is not an option, given that
performPoll might run several times per second by system server
callers.
This change adds a reason code to the performPoll method and
a helper class to log performPoll events. This will allow us
to track down the caller of performPoll and better understand
why it is being called frequently or causing performance
problems.
This change also introduces a PollEvent object, which can be
used to put related information in subsequent changes when
logging.
Sample output:
Poll counts per reason:
DUMPSYS: 0
FORCE_UPDATE: 6
GLOBAL_ALERT: 0
NETWORK_STATUS_CHANGED: 3
OPEN_SESSION: 0
PERIODIC: 1
RAT_CHANGED: 1
REG_CALLBACK: 3
REMOVE_UIDS: 0
UPSTREAM_CHANGED: 1
Recent poll events:
2023-10-04T02:47:12.937370 - Poll(flags=3, PollEvent{reason=REG_CALLBACK})
2023-10-04T02:47:13.178559 - Poll(flags=3, PollEvent{reason=FORCE_UPDATE})
Test: manual
Test: atest FrameworksNetTests:android.net.connectivity.com.android.server.net.NetworkStatsEventLoggerTest
Test: atest FrameworksNetTests:android.net.connectivity.com.android.server.net.NetworkStatsServiceTest
Bug: 301141955
Change-Id: Ic6d0543ccdef12493385f45200199854b841a26b
* changes:
Add toString for more consistent logging to IaPrefixOption
Create IpPrefix in IaPrefixOption constructor
Add @Computed annotation to Struct
Remove check that preferred lifetime >= t2
Remove IA prefix option option-code from public constructor
Patrick Rohrauthored and Gerrit Code Reviewcommited 2023-10-25 02:14:30 +00:00
The constructor is invoked during Struct#parse() at which point we might
as well try to create the IpPrefix object which will be used later. If
this fails for some reason (UnknownHostException or ClassCastException),
the prefix is bogus and should not be used.
Test: TH
Change-Id: I7e0090735022d017ba11aa85266ba19e65807da7
@Computed will ignore "computed" fields from struct parsing. Fields
cannot both be annotated by @Computed and @Field (though there
currently is no check for this).
Test: builds
Change-Id: I29fe506c5e3be4cc50fdaf3a07fc5e922111165b
For a single address, DHCPv6 won't work well (i.e. potentially lose provisioning) if
the preferred lifetime is lower than t2. However, this might be a valid
scenario in a renumbering event in the presence of multiple prefixes.
Additionally, in DHCPv6 it is up to the server to configure the client
correctly.
Test: TH
Change-Id: I37b2a0b89deda85b3f7be29c8f02f685aa6e65f8
Mark Chien
authored and
Gerrit Code Review
commited 
Maciej Żenczykowski
authored and
Ken Chen
authored and
Jean Chalard
authored and
KH Shi
authored and
Patrick Rohr
authored and
Treehugger Robot
authored and
Krzysztof Kosiński
authored and
Kangping Dong
authored and
Motomu Utsumi
commited 
Junyu Lai
commited