Age | Commit message (Collapse) | Author |
|
Change-Id: Ibbd3d7792162bcdbaaffc889159d5718c945d901
|
|
The upstream commit (8fe45924387be6b5c1be59a7eb330790c61d5d10 bpf:
map_get_next_key to return first key on NULL) introduce the support for
bpf maps to return the first key of the map when receive a NULL key from
userspace. This change will test this functionality in supported kernels
to make sure they have the correct backport.
Test: ./bpf_test.py
Bug: 30950746
Change-Id: I6a497a3ccb518ebb037bfa447f827d2715271291
Merged-In: I6a497a3ccb518ebb037bfa447f827d2715271291
(cherry picked from aosp commit 493cfb50cb18677628f0ae62dd8a14036a8eec0c)
|
|
Change-Id: I5bc6ea8ad0c8d92ecf2481bf65a50ffb1a96a2d5
|
|
The intent of PollOnCloseTest is to ensure that SOCK_DESTROY
behaves in the same way with respect to poll() as receiving a TCP
RST packet would.
Unfortunately, due to a race inside the kernel, the return values
of poll() in these circumstances is non-deterministic. The race
when processing a TCP RST has been fixed, but the one when
processing SOCK_DESTROY has not.
Ensure that the test does not insist on specific return values on
kernels that do not guarantee them.
Also ensure that any assertion failures in SocketExceptionThread
are passed on to the main thread and properly cause the test to
fail.
Bug: 73654768
Test: sock_diag_test passes 100 times in a row on taimen PPR1.180418.002.
Change-Id: I509e23a67be842d3d112a165f1439d29c699b01f
Merged-In: I509e23a67be842d3d112a165f1439d29c699b01f
|
|
Change-Id: Ic12b1d341d24ff436503f46f40eb30aea29ae720
|
|
|
|
Change-Id: I9663b95c685b1ff3ae4d2da48e3a4e2d046c9765
|
|
This test checks that an existing VTI(6) can be
updated using the RTM_NEWLINK without the EXCL
flag.
Bug: 72532478
Test: aosp/kernel/tests/net/test/run_net_test.sh
Test: kernel_net_tests (on wahoo/walleye)
Merged-In: Iecb7bb7f070dfbdd098dbb773d18ca21f23a42da
Change-Id: Iecb7bb7f070dfbdd098dbb773d18ca21f23a42da
(cherry picked from commit 3f132dc23b080bd6c1951606346ade7e1dbadd41)
|
|
With the backport of 00bc0ef, 3.18 kernels no longer hang on the
xfrm algorithm tests. This patch re-enables these tests for all 3.18
kernels.
Bug: 63589559
Test: All kernel tests run, passing.
Merged-In: I27b7e463a5566b1211ef6248f9239a52ff39d50d
Change-Id: I27b7e463a5566b1211ef6248f9239a52ff39d50d
(cherry picked from commit 2ab89c98a1431747fd2c1a2912732c26ba4c1124)
|
|
Lingering TCP sockets appear to be causing testUpdatePolicy to fail due
to policies attached to TCP sockets in TIME_WAIT. This patch sets
SO_LINGER on the algorithm tests' sockets, forcing them to shutdown
cleanly, and synchronously.
Bug: 74071662
Test: Ran on common 3.18, 4.4 with parallel_tests.sh 20 10
Merged-In: Ie9ab9b7672cc2ce8309ed871eeee09ec8128aac9
Change-Id: Ie9ab9b7672cc2ce8309ed871eeee09ec8128aac9
(cherry picked from commit f7699431b8717d77c24b517dfb8ef43065874698)
|
|
Change-Id: I72da158352fa5e3912a490bf8fa3feef853c3971
|
|
am: a0309c3a69
Change-Id: Ifa510c57ed828af602bafcfe676aec3d6d645e1e
|
|
am: a1ea097b76
Change-Id: Ia83b67d56aad47ade30b60ba15a7c13df9d83183
|
|
am: 450b10755c
Change-Id: I59fdf513db533806c05d89c52ac547dbae27a7e2
|
|
Add two test to verify the kernel patch of creating and using bpf map in
read only and write only mode.
Test: new test pass on 4.9 kernel and above.
Bug: 30950746
Change-Id: I74e94a9ef4b4e23bb4f9e00340ecf701e1f13060
|
|
Change-Id: I232d22ad2ebd60959e041591a6d00c462e26033e
|
|
Change-Id: Ideba30485d4bbdee840f96f6b5d0983ddef748dc
|
|
am: fbf22f11da
Change-Id: Iea97aff5175fe8c5ac1a02a96d24352c2668d7c1
|
|
am: 3ef7d06c05
Change-Id: I6c176290503d944d9d0d73f3b0e3bdb9f4b1d593
|
|
am: 8544ec3f40
Change-Id: Ief1b942270627633725b621a508f3951b1f4b1a3
|
|
* changes:
Make sock_diag_test.py pass on 4.9 devices that don't have SCTP.
Make BPF tests pass on device.
|
|
The test code currently checks that we can see marks set on SCTP
sockets. This doesn't work on devices that don't have SCTP. This
has never been a problem on VM tests because the VM tests enable
the SCTP config option.
Bug: 36510518
Test: sock_diag_test.py passes on android-4.9
Test: sock_diag tests pass on 4.9 device
Change-Id: I82fd77939be549a0b2c6a2ab0395b66e7b2bdc80
|
|
am: a8d6cac035
Change-Id: I5aaeb9fe791a7b72c329da64a2adf593a10bf191
|
|
am: 1440dfc47d
Change-Id: If7e92b5a44eb1ca0d5fd48cd22377ed2a4974e6e
|
|
am: 8085ed76b3
Change-Id: I2a58f51c31869d54b981ae5c0ab1a42ee8322c45
|
|
1. Add __NR_bpf constants for armv8l and aarch64.
2. Move the temporary cgroup BPF mount from /tmp (which does not
exist on device) to a temporary directory.
3. Use proper ctypes functions to pass pointers.
Bug: 36510518
Test: all_tests.sh passes on android-4.9
Test: 32-bit and 64-bit BPF tests pass on 4.9 device
Change-Id: Id3cca05d8ea547aeaa5cfd1d2d88b22115fe432d
|
|
Change-Id: I48ce6ad488045ddb294b547bc70bff27c0b840bc
|
|
XfrmVtiTest now creates one IPv4 and one IPv6 tunnel per netid.
Similar to MultinetworkBaseTest, the tunnels are created in
setUpClass and torn down in tearDownClass, so all the tests are
always run with all the VTIs present.
This provides the basic infrastructure to perform more in-depth
testing of the VTI code in a realistic environment.
Bug: 70371070
Test: xfrm tests pass on android-4.9
Change-Id: I295c3af6626b0ced98975bbb67c2592ab9a8f455
|
|
4b563d2e8e
am: 7b24217ff8
Change-Id: Idf1b70083303f525078b5c59be59881cc0b25f81
|
|
am: 4b563d2e8e
Change-Id: I98c0d246f959075f5ee4e3b19e9241eabedfb5c4
|
|
am: 31c8de4f14
Change-Id: I283b646977013052b782cd89d8816d8d82e0d21a
|
|
This class encapsulates both the knowledge needed to create a
VTI interface and the state that it holds. This will be useful
in future changes to store multiple VTIs in the same test.
Test: xfrm tests pass on android-4.9
Change-Id: I5b047a41e384b94b7cf50546026b32081f5d00bd
|
|
Change-Id: I24ef3d5ef1be4464c32a30348705c768c2c8e0c6
|
|
79aeff1d35
am: 04d0f44b7f
Change-Id: I753586eedee46753b92fb368b7074cbd31c207fa
|
|
am: b5d42332d5
Change-Id: I378ba37d86d3f3bc9e0c61cd17e1b985da0fdffe
|
|
am: 79aeff1d35
Change-Id: I8c27a293725c08de2755dfb2140cb80b0c39df7b
|
|
am: 4431616b6b
Change-Id: Ie22f726c8bb4caf801243638ca7d7deaddfbb14b
|
|
am: 3a38134704
Change-Id: I783c83782088fbb7a8e9d017f081eeb51fb2e08d
|
|
am: a54e221f5d
Change-Id: I855b2f91f41ff4568c7647a6da972646af4aae82
|
|
|
|
Exit code should be:
0 on success
non-zero on failure
Bug: 36510518
Test: m -j32 kernel_net_tests && adb sync data && adb shell
/data/nativetest64/kernel_net_tests/kernel_net_tests
Change-Id: Ief32d887828adb4a7e63454b79aca7d24a0a0567
|
|
3491aa0df7
am: 0e83fe9786
Change-Id: I57f4fcb446499a88950873c2bbfa42626bd27d2e
|
|
am: 3dedef87e5
Change-Id: I40082398c782fbdfb72acca5f5bec80d71a9743c
|
|
am: 3491aa0df7
Change-Id: I87bed1a5209ddeef1f1ae93a838bd1bc422124f1
|
|
am: 7a5cd50e70
Change-Id: Ib8307044d6c2fee164f84d47c6e7905e4e876fdb
|
|
am: e37079467e
Change-Id: I85704fba9836b767107cb007c22d8469322aa30c
|
|
am: 54e55c9cf1
Change-Id: I2919458b4be0ae4956c00bff306f07f105709978
|
|
am: eec7171ecf
Change-Id: I7021815bf1901eb7d4e517e1a0c31137e95e9de4
|
|
Bug: 70371070
Test: xfrm_tunnel_test.py passes on android-4.9
Change-Id: I3a313a0c158efd084cb5601c0ae3c0999fc55467
|
|
This will allow tests to set up XFRM state and policies in
setUpClass instead of in setUp. This speeds up running the tests
because setting up iptables rules can take a couple of seconds.
This is also useful when running on device, where blowing away
XFRM state in tests could be disruptive to other system
functionality.
Bug: 70371070
Test: xfrm tests pass on android-4.9
Change-Id: I6246c172368fd65a815175ecf58b53965de1bcfd
|