The WHATWG URL parsing algorithm [1] used by browsers says that for
"special" URL schemes (which is basically all commonly-used
hierarchical schemes, including http, https, ftp, and file), the host
portion ends if a \ character is seen, whereas this class previously
continued to consider characters part of the hostname. This meant
that a malicious URL could be seen as having a "safe" host when viewed
by an app but navigate to a different host when passed to a browser.
[1] https://url.spec.whatwg.org/#host-state
Bug: 71360761
Test: vogar frameworks/base/core/tests/coretests/src/android/net/UriTest.java (on NYC branch)
Test: cts -m CtsNetTestCases (on NYC branch)
Change-Id: Id53f7054d1be8d59bbcc7e219159e59a2425106e
Malformed authority segments can currently cause the parser to produce
a hostname that doesn't match the hostname produced by the WHATWG URL
parsing algorithm* used by browsers, which means that a URL could be seen
as having a "safe" host when checked by an Android app but actually visit
a different host when passed to a browser. The WHATWG URL parsing
algorithm always produces a hostname based on the last @ in the authority
segment, so we do the same.
* https://url.spec.whatwg.org/#authority-state resets the "buffer", which
is being used to build up the host name, each time an @ is found, so it
has the effect of using the content between the final @ and the end
of the authority section as the hostname.
Bug: 68341964
Test: vogar android.net.UriTest (on NYC branch)
Test: cts -m CtsNetTestCases (on NYC branch)
Change-Id: Idca79f35a886de042c94d6ab66787c2e98ac8376
MemoryIntArray was using the size of the undelying
ashmem region to mmap the data but the ashmem size
can be changed until the former is memory mapped.
Since we use the ashmem region size for boundary
checking and memory unmapping if it does not match
the size used while mapping an attacker can force
the system to unmap memory or to access undefined
memory and crash.
Also we were passing the memory address where the
ashmem region is mapped in the owner process to
support cases where the client can pass back the
MemoryIntArray instance. This allows an attacker
to put invalid address and cause arbitrary memory
to be freed.
Now we no longer support passing back the instance
to the owner process (the passed back instance is
read only), so no need to pass the memory adress
of the owner's mapping, thus not allowing freeing
arbitrary memory.
Further, we now check the memory mapped size against
the size of the underlying ashmem region after we do
the memory mapping (to fix the ahsmem size) and if
an attacker changed the size under us we throw.
Tests: Updated the tests and they pass.
bug:33039926
bug:33042690
Change-Id: Ibf56827209a9b791aa83ae679219baf829ffc2ac
MemoryIntArray was using the size of the undelying
ashmem region to mmap the data but the ashmem size
can be changed until the former is memory mapped.
Since we use the ashmem region size for boundary
checking and memory unmapping if it does not match
the size used while mapping an attacker can force
the system to unmap memory or to access undefined
memory and crash.
Also we were passing the memory address where the
ashmem region is mapped in the owner process to
support cases where the client can pass back the
MemoryIntArray instance. This allows an attacker
to put invalid address and cause arbitrary memory
to be freed.
Now we no longer support passing back the instance
to the owner process (the passed back instance is
read only), so no need to pass the memory adress
of the owner's mapping, thus not allowing freeing
arbitrary memory.
Further, we now check the memory mapped size against
the size of the underlying ashmem region after we do
the memory mapping (to fix the ahsmem size) and if
an attacker changed the size under us we throw.
Tests: Updated the tests and they pass.
bug:33039926
bug:33042690
Change-Id: I1004579181ff7a223ef659e85c46100c47ab2409
MemoryIntArray was using the size of the undelying
ashmem region to mmap the data but the ashmem size
can be changed until the former is memory mapped.
Since we use the ashmem region size for boundary
checking and memory unmapping if it does not match
the size used while mapping an attacker can force
the system to unmap memory or to access undefined
memory and crash.
Also we were passing the memory address where the
ashmem region is mapped in the owner process to
support cases where the client can pass back the
MemoryIntArray instance. This allows an attacker
to put invalid address and cause arbitrary memory
to be freed.
Now we no longer support passing back the instance
to the owner process (the passed back instance is
read only), so no need to pass the memory adress
of the owner's mapping, thus not allowing freeing
arbitrary memory.
Further, we now check the memory mapped size against
the size of the underlying ashmem region after we do
the memory mapping (to fix the ahsmem size) and if
an attacker changed the size under us we throw.
Tests: Updated the tests and they pass.
bug:33039926
bug:33042690
Change-Id: Id7f0e8a4c861b0b9fa796767e0c22d96633b14d1
MemoryIntArray was using the size of the undelying
ashmem region to mmap the data but the ashmem size
can be changed until the former is memory mapped.
Since we use the ashmem region size for boundary
checking and memory unmapping if it does not match
the size used while mapping an attacker can force
the system to unmap memory or to access undefined
memory and crash.
Also we were passing the memory address where the
ashmem region is mapped in the owner process to
support cases where the client can pass back the
MemoryIntArray instance. This allows an attacker
to put invalid address and cause arbitrary memory
to be freed.
Now we no longer support passing back the instance
to the owner process (the passed back instance is
read only), so no need to pass the memory adress
of the owner's mapping, thus not allowing freeing
arbitrary memory.
Further, we now check the memory mapped size against
the size of the underlying ashmem region after we do
the memory mapping (to fix the ahsmem size) and if
an attacker changed the size under us we throw.
Tests: Updated the tests and they pass.
bug:33039926
bug:33042690
Change-Id: Ie267646eb88014034fbd048d7a9bc273420c7eff
Properly account for VPN apps that make heavy use of the tun
interface. Prior to this change a VPN app could be incorrectly charged
for more data than it actually used if it sent more traffic through
the tun interface than the underlying interface.
This change excludes VPN app traffic on the tun interface from the
adjustment pool and doesn't redistribute traffic to the VPN app.
Instead all of the redistributed traffic is deducted from the VPN app
which effectively represents any overhead incurred by the VPN app.
BUG: 30557871
Change-Id: I62a75a0c0c0111e052b7903baa9f5d6d94ef57fd
When we create a file that already exists, we try attaching a suffix
like "(1)" to the filename to avoid the conflict. The newly added
rename method should do the same, since developers may not have
access to delete the conflicting file.
Test: boots, rename via UI, new unit tests
Bug: 31545404
Change-Id: Ie397eebb0fbf98cf079eee3bbbb6c6b7ca627d91
LoaderManagers configure their host callback lazily as their
associated fragment is brought up through its lifecycle states. In the
case of fragments on the fragment back stack this could happen very
late, if at all. As a LoaderManager's host callback references the
host Activity, this means that a LoaderManager could keep a destroyed
Activity reference alive.
Update the host callbacks of all LoaderManagers eagerly during the
restore non-configuration instance phase.
Bug: 30653222
Test: core/tests/coretests/src/android/app/LoaderLifecycleTest.java
Change-Id: I5d2b81daae5e7cae429fcf4934e64b3ce281140c
This change adds a new pathMatch attribute to the intent-filter tag
that supports a subset of regular expression syntax. Supported
tokens include dot (.) and set ([]) and supported modifiers include
zero-or-more (*), one-or-more(+) and ranges ({}).
Change-Id: I2ec3bc8a9eb3e94d21caf34a1ed0f3fd3fb33a35
Bug: 30805203
The summary is supposed to just hold enough data to continue counting
once the device has reset. Since kernel stats reset when the device
resets, and the first update is ignored to account for soft resets where the
kernel continues running, SamplingTimer should not be recording the last value
it saw from /proc/wakelocks in the summary.
Bug:30575302
Change-Id: Ic193bc5af9a0ede514e3abc8146523d7316c47d3
Tests that long stack traces and exception messages are
truncated to a reasonable size when a CrashReport is generated.
Bug: 29918978
Change-Id: I4d8202ddca850a72e0853e4064d5690965ff371c
