| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
| |
Test: built, flashed, booted
system/netd/tests/runtests.sh passes
Change-Id: Ibff8b050915c65e44cd37bb835943b8582fc003a
|
| |
|
|
|
|
| |
Bug: 78370064
Test: lunch aosp_walleye-userdebug && cd system/netd && mma
Change-Id: I495f7cdc3f2aab26947f10041fc559700367f2ea
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently, when applying a cleartext policy to a UID,
StrictController will attempt to delete all possible policies
that might previously have applied to this UID. Because only
two of these rules can exist at any given time, at least one
of these deletes is guaranteed to fail, causing the whole
operation to fail.
Instead of adding a log or reject rule for every UID, add a
rule that sends that UID to its own chain which then contains
the log or reject rule. That way, deleting the previous policy
only requires deleting the chain, which is something we know
exists.
(cherry picked from commit 81e02f21424ebda9a314c7b9fbc31cfe2a99444f)
Bug: 64988066
Test: netd_{unit,integration}_test pass
Test: android.os.cts.StrictModeTest passes
Change-Id: Ic9d66220a65f2ce9510c4194e7b874d3d5dca5d7
|
| |
|
|
|
|
| |
Bug: 28362720
Test: unit tests pass
Change-Id: I8a4d2b8ea66799c6c3205b00f04ee1999fc7c68b
|
| |
|
|
|
|
|
| |
Test: bullead builds and boots with no iptables errors
Test: netd_{unit,integration}_test pass
Bug: 32323979
Change-Id: I33ad04ee8f0562bcd4e14046352c934cd2039a5d
|
| |
|
|
|
| |
Bug: 21725996
Change-Id: I2c049a934189f3c87ee15f052abc07d35814f0c9
|
| |
|
|
|
|
| |
Bug: 21725996
Bug: 25691379
Change-Id: I24b838161eaf98dede2ae897157ba42414fc926f
|
| |
|
|
|
|
|
|
| |
MARK/CONNMARK values/tags are shared accross all controllers because
of the way the firewall works. To avoid accidental clashes, it's best
to store the values used in a central place.
Change-Id: I76aaba38cba6554704a5635b1e7297a144e6e2ff
|
|
|
Introduces new module that provides network-related features for
the StrictMode developer API. The first feature offers to detect
sockets sending data not wrapped inside a layer of SSL/TLS
encryption.
This carefully only adds overhead to UIDs that have requested
detection, and it uses CONNMARK to quickly accept/reject packets
from streams that have already been inspected. Detection is done
by looking for a well-known TLS handshake header; it's not future
proof, but it's a good start. Handles both IPv4 and IPv6.
When requested, we also log the triggering packet through NFLOG and
back up to the framework to aid investigation.
Bug: 18335678
Change-Id: Ie8fab785139dfb55a71b6dc7a0f3c75a8408224b
|