| d3782f304acee54273f551c52f547539c8f4a78b |
|
05-Apr-2018 |
Alex Light <allight@google.com> |
Prevent deadlock calling transformation functions in ClassLoad callback.
If an agent called RetransformClasses or RedefineClasses from the
ClassLoad event callback and the class that caused the event is
included the thread would deadlock. This happened because the
verification code would try to wait for the class to be linked but
since the current thread is responsible for doing that this will never
happen. To prevent this from happening we make those functions simply
return JVMTI_ERROR_INTERNAL if they are called on the same thread the
class is being loaded on.
In order to test this we needed to modify the test helper code to keep
track of the current jvmtiEventCallbacks state.
Bug: 77652488
Test: ./test.py --host -j50
(cherry picked from commit 9e7859cb2449d6150b97c7d8ff9646a5405737f8)
Merged-In: I0a196b999a08ec3bf9cdf98357e223f89fdcd666
Change-Id: I0a196b999a08ec3bf9cdf98357e223f89fdcd666
|
| 12ee56bcddfd36e254aaf855a80e0fae2b91c97a |
|
12-Feb-2018 |
Alex Light <allight@google.com> |
Fix incorrect condition in TransformationFaultHandler
We were incorrectly checking if an iterator was
uninitialized_class_definitions_.end() instead of
initialized_class_definitions_.end(). This caused several tests to
fail when run with redefine-stress because they would cause the
faulting code to resume incorrectly.
Test: ./test.py --host -j50
Test: ./test.py --host --redefine-stress -j50
Bug: 73177368
Change-Id: I9ca86aa4e30f342d73545d1a417bab8a0e817251
|
| fe2a39d1a4297b3eb3bbf5d5af3eae8172b8c4db |
|
05-Feb-2018 |
Alex Light <allight@google.com> |
Revert "Temporarily disable lazy class-file-load-hook"
An ASAN failure was caused by a rare interaction between the
fault-handler and updating the internal book-keeping for the
ClassFileLoadHook. When we update the current definition for the
ClassFileLoadHook we check to see if the definition has actually
changed. This causes a read from the previous dex-file buffer. If the
buffer is lazy and the agent never accessed it (very rare outside of
testing) the SEGV will occur within art code in the kRunnable state.
This caused the runtime to think the SEGV was coming from java code
and perform some illegal reads prior to figuring out that the SEGV did
not come from java.
To fix this we simply make sure all the code that can cause the SEGV
occurs while the runtime is in kNative state.
We add an assert for this in the fault handler itself.
Reason for revert: Fixed issue causing ASAN failure with lazy dex-file
load hook.
This reverts commit 37977705dc6cfa3e77299ae97675f80b59f2fcf0.
Bug: 72909916
Test: ./test/testrunner/run-_build_test_target.py -j50 art-asan
Change-Id: I8b2d2fda8092a379e0019db15ec360a5a63d8b3f
|
| ca97ada4ec14373242de2452a6d0c16804007d66 |
|
02-Feb-2018 |
Alex Light <allight@google.com> |
Change ClassFileLoadHook to lazily compute dex file
Creating a dex file from the quickened or compact-dex'd internal
format for calling the JVMTI ClassFileLoadHook is quite expensive.
This meant that agents could not generally listen for this event
without causing unacceptable performance problems.
Since agents will generally not touch the buffer, needing to
instrument only a handful of classes we will fix this problem by doing
the de-quickening lazily. This is done by mmaping an empty buffer with
PROT_NONE and then filling it in when the program has a SEGV in the
appropriate address range. This should improve the performance of any
agent that listens for the ClassFileLoadHook but does not commonly do
anything to the buffer.
This does have the disadvantage that we can no longer ensure that the
buffer size we pass down in class_data_len might no longer be fully
accurate. Some agents that assert that class_data_len is exactly the
same as the dex-file will need to be updated.
Bug: 72402467
Bug: 72064989
Test: ./test.py --host -j50
Test: ./test.py --host --redefine-stress -j50
Change-Id: I39354837f1417ae10a57c5b42cbb4f38f8a563dc
|
| 0225f8e2939a9340cb7dcebfcfe7996a2bd9bce9 |
|
31-Jan-2018 |
David Sehr <sehr@google.com> |
Revert "Revert "Make libdexfile build independent of runtime dir""
This reverts commit 787784f9effb126b5d0d3dc97d544c4a477b5daf.
Reason for revert: Bot configuration issue.
Change-Id: I6a10bb4a9571f89c7e4dd095f9157e830a44e2de
Bug: 22322814
Test: make -j 50 checkbuild
|
| 64e4c145e9301615a339a0981b1bc34fbe78c2f9 |
|
30-Jan-2018 |
Alex Light <allight@google.com> |
Refactor JVMTI ClassFileLoadHook handling
Make initial load ClassFileLoadHook use the same helper structures as
the retransform path. This makes the system simpler and creates a
single point where changes to the ClassFileLoadHook handling can be
done.
Test: ./test.py --host -j50
Test: ./test.py --redefine-stress --host -j50
Bug: 72064989
Change-Id: I7df9fed267fea497d9efd4f4214506237e63339f
|
| 787784f9effb126b5d0d3dc97d544c4a477b5daf |
|
30-Jan-2018 |
David Sehr <sehr@google.com> |
Revert "Make libdexfile build independent of runtime dir"
This reverts commit b40b7e73469339a6b667b4a2e2b8690112a74dc9.
Reason for revert: on device libdexfile.so missing
Change-Id: I9bd61a98bef870400580e8c991cb061d3f57fa72
|
| b40b7e73469339a6b667b4a2e2b8690112a74dc9 |
|
26-Jan-2018 |
David Sehr <sehr@google.com> |
Make libdexfile build independent of runtime dir
Remove libdexfile's dependency on utils.cc and move utf.cc into
/dex. Remove libdexfile's constituent sources from libart and
use libdexfile wherever libart is. Also remove some ART-specific
interfaces. Libdexfile's tests remain to be converted, plus
moving the files to a new directory peer to runtime/.
Bug: 22322814
Test: make -j 50 test-art-host
Change-Id: Ifaf695216e4a0e43d3aa377984d933f7a2a243c2
|
| 9e734c7ab4599d7747a05db0dc73c7b668cb6683 |
|
05-Jan-2018 |
David Sehr <sehr@google.com> |
Create dex subdirectory
Move all the DexFile related source to a common subdirectory dex/ of
runtime.
Bug: 71361973
Test: make -j 50 test-art-host
Change-Id: I59e984ed660b93e0776556308be3d653722f5223
|
| 06c42a571358b5e5adb69104b183af8f32f4c07d |
|
26-Jul-2017 |
Andreas Gampe <agampe@google.com> |
ART: Move openjdkjvmti to art/
Move libopenjdkjvmti out of the runtime directory. Let's not
pollute the runtime library.
Test: m test-art-host
Change-Id: Idb6b9cebcd61777bd3200437a2ae584a63a4a341
|