While studying the implementation of POSIX pthread_rwlock* functions, I noticed
that two functions were marked __always_inline twice. "They must really mean it
this time."
Also add back `inline` keyword to one other usage of __always_inline to be
consistent with other uses of __always_inline throughout the codebase.
Change-Id: Ibf9eaed5fc9fd03afcdd969cff82dec71a8ce30f
Adding the new option log_allocator_stats_on_signal that will call
mallopt(M_LOG_STATS, 0), when a signal is sent to a process.
This call does not happen in the signal handler, but a variable
is set to trigger the mallopt call. The next time any allocation
call occurs, the mallopt call will be made.
Also, includes new unit tests for the new config option.
Test: All unit tests pass.
Test: Enabling the new feature and stop/start and sending the signal to
Test: a process and observing the log contains the memory stats.
Test: Did the above for a scudo and jemalloc based device.
Change-Id: Idcf6fca74036efb065b9c4cc037aaf9bcf0f139e
Since it doesn't matter if these calls take a little longer than
before, use the more thorough but slightly longer purge mechanism.
Test: Unit tests pass.
Change-Id: Ifab7166a9682a13231746b78717d52673d13be1b
Some operations wind up allocating then freeing a significant
amount of memory. So after those operations, do a purge so that
the RSS of the process is not artificially inflated.
Bug: 262321164
Test: Ran unit tests.
Test: Verified after this change, the RSS does not go up after running
Test: am dumpheap -n <PID>.
Change-Id: I08477f8ce12c06fd2a068f536a81f4a577d619e2
We swap the 2nd and 3rd arguments to the CallocEntry constructor
to match the order in the cpp file, and match the C calloc convention.
We also fix an invocation of this constructor.
Change-Id: Iebe16d82a74459e5e957c1d9e2cc1aebb15150d0
Test: TreeHugger
With timestamps, we are able to tell the details of allocator
performance such as the average time for malloc() in different size
class, the potential contention time by examing the overlap between
operations, .etc.
Not all malloc et al. operations are recorded with timestamp. Only
operations relates to memory usage change will have them.
Test: All unit tests pass.
Change-Id: I5c2016246a6f10b221387001bb44778969bb26ae
Add backtrace_size for only backtracing a single size.
Add backtrace_min_size to set the minimum value of size to backtrace.
Add backtrace_max_size to set the maximum value of size to backtrace.
Documented the new options.
Test: New unit test pass.
Change-Id: I1a773737910cd4bc2af9546547b3a2740bbcb22b
The new option is named check_unreachable_on_signal. It is meant
to duplicate dumpsys meminfo --unreachable <PID> for non-java
processes. When enabled, a user can send a signal to a process
which will trigger the unreachable check on the next allocation
call.
Added new unit tests.
Test: New unit tests pass.
Test: Enabled for the entire system, then dumped on the netd
Test: process and also system_server.
Change-Id: I73561b408a947a11ce21a211b065d59fcc39097b
The libmemunreachable library looks through memory to determine
if pointers are leaked. Unfortunately, the malloc debug code
stores the original pointer in data structures, so it looks like
pointers are still in use. The fix is to mangle the pointers
stored in memory so that it doesn't trick the library into thinking
they are live.
Test: All unit/system tests pass.
Test: Ran libmemunreachable and verified leaks show up.
Change-Id: Ic40a0a5ae73857cde936fd76895d88829686a643
Modify libfdtrack to use the normal Unwinder object. In addition,
update the libfdtrack so that it doesn't record frames in
libfdtrack.so rather than skipping frames it thinks will be in
the library.
Modify the malloc debug code to use the normal Unwinder object.
Bug: 120606663
Test: All unit tests pass.
Change-Id: I3c9612dd10e62389e6219e68045ee87f7b2625f5
When the main thread is exiting, the code deleted the g_debug global
pointer and destroys the disable pthread key. Unfortunately, if
malloc debug was enabled in a way that requires a header for the pointer,
any frees that occur after the main thread is torn down result in calls
to the underlying allocator with bad pointers.
To avoid this, don't delete the g_debug pointer and don't destroy the
disable pthread key.
Added a new system test that allocates a lot of pointers and frees them
after letting the main thread finish.
Also, fix one test that can fail sporadically due to a lack of unwinding
information on arm32.
Bug: 189541929
Test: Passes new system tests.
Change-Id: I1cfe868987a8f0dc880a5b65de6709f44a5f1988
When malloc debug is enabled, using libbacktrace to unwind can
result in a deadlock. This happens when an unwind of a thread
is occuring which triggers a signal to be sent to that thread. If
that thread is interrupted while a malloc debug function is
executing and owns a lock, that thread is then stuck in the signal
handler. Then the original unwinding thread attempts to do an
allocation and gets stuck waiting for the same malloc debug lock.
This is not a complete deadlock since the unwinder has timeouts,
but it results in truncated unwinds that take at least five
seconds to complete.
Only the backtrace signals needs to be blocked because it is the only
known signal that will result in a thread being paused in a signal
handler.
Also, added a named signal in the reserved signal list for the
special bionic backtrace signal.
Bug: 150833265
Test: New unit tests pass with fix, fail without fix.
Change-Id: If3e41f092ebd40ce62a59ef51d636a91bc31ed80
This patch introduces tagged pointers to bionic. We add a static tag to
all pointers on arm64 compatible platforms (needs requisite
top-byte-ignore hardware feature and relevant kernel patches).
We dynamically detect TBI-compatible devices (a device with the TBI feature and
kernel support) at process start time, and insert an implementation-dependent
tag into the top byte of the pointer for all heap allocations. We then check
that the tag has not been truncated when deallocating the memory.
If an application incorrectly writes to the top byte of the pointer, we
terminate the process at time of detection. This will allow MTE-incompatible
applications to be caught early.
Bug: 135754954
Bug: 147147490
Test: cd bionic && atest .
Change-Id: Ie424325ba1e3c4443040ac265aeaa28d9e405d28
This reverts commit 43d5f9d4dd.
Bug: 135754954
Bug: 147147490
Exempt-From-Owner-Approval: clean revert
Reason for revert: Breaks ART gtest, see:
https://ci.chromium.org/p/art/builders/ci/angler-armv8-non-gen-cc/561
The crash happens on mprotect of a page, the test crashes with ENOMEM.
Change-Id: I52eea1abbfaf8d8e2226f92d30aa55aba3810528
This patch introduces tagged pointers to bionic. We add a static tag to
all pointers on arm64 compatible platforms (needs requisite
top-byte-ignore hardware feature and relevant kernel patches).
We dynamically detect TBI-compatible devices (a device with the TBI feature and
kernel support) at process start time, and insert an implementation-dependent
tag into the top byte of the pointer for all heap allocations. We then check
that the tag has not been truncated when deallocating the memory.
If an application incorrectly writes to the top byte of the pointer, we
terminate the process at time of detection. This will allow MTE-incompatible
applications to be caught early.
Bug: 135754954
Bug: 147147490
Test: cd bionic && atest .
Change-Id: I6e5b809fc81f55dd517f845eaf20f3c0ebd4d86e
I have no idea why I used the iterate name internally which is
completely unlike every other function name. Change this to match
everyone else so that it's now malloc_iterate everywhere.
This is probably the last chance to change this before mainline
modules begin, so make everything consistent.
Test: Compiles, unit tests passes.
Change-Id: I56d293377fa0fe1a3dc3dd85d6432f877cc2003c
When using a FILE object for some malloc debug functions, calling
fprintf will trigger an allocation to be put in the object. The problem
is that these allocations were not allocated by the malloc debug
wrapper and they get freed during the fclose as if they are malloc
debug allocation. In most cases, the code will detect the bad pointer
and leak the memory, but it might also cause a crash.
The fix is to avoid using fprintf so that no allocations are made
in the object that survive and need to be freed in the fclose call.
Change the MallocXmlElem.h to use a file decsriptor not a FILE object.
Add new unit and system tests to detect this case.
Bug: 143742907
Test: Ran unit and system tests.
Test: Ran bionic unit tests.
Change-Id: I524392de822a29483aa5be8f14c680e70033eba2
Bug: 123312263
Test: Verified with asan enabled, malloc debug does not initialize.
Test: Ran tests on non-asan build and verify they pass.
Change-Id: I3c37c170bf6c1de42740972f2113ae991351d931
I wrote a new unit test that would fail on the old version of the
code.
On a walleye big cpu, this costs about 40ns-50ns (going from ~430ns to ~480ns).
I think this is an acceptable performance degradation.
Bug: 131867816
Test: New unit tests pass.
Change-Id: I4c0f4373fb0694bf29c3824dbb1224a8a17e211e
Remove this global variable and change the setting of it to non-zero
to a call to android_mallopt.
In addition, change the initialize function to use pass a bool* instead of
int*.
Bug: 130028357
Test: Ran malloc_debug/malloc_hooks/perfetto tests.
Change-Id: I20d382bdeaaf38aac6b9dcabea5b3dfab3c945f6
Merged-In: I20d382bdeaaf38aac6b9dcabea5b3dfab3c945f6
(cherry picked from commit 5225b342f0)
Add a new option verbose for malloc debug that is not enabled by default.
This disables all of the info log messages. It turns out these log
messages can add a measurable amount of time and can change the boot up.
Bug: 129239269
Test: Adjusted unit tests pass.
Test: Verified no messages unless verbose option used.
Change-Id: I805cb7c8ecb44de88119574e59d784877cacc383
malloc_info needs to be per native allocator, but the code treated it
like a global function that doesn't depend on the native memory allocator.
Update malloc debug to dump the actual pointers that it has been tracking.
Test: bionic-unit-tests pass.
Test: malloc debug tests pass.
Test: malloc hook tests pass.
Change-Id: I3b0d4d748489dd84c16d16933479dc8b8d79013e
Merged-In: I3b0d4d748489dd84c16d16933479dc8b8d79013e
(cherry picked from commit a3656a98b1)
Jemalloc does not verify that the size parameter is a multiple of
alignment. Fix this since it only went into P.
Fix the unit tests, and fix malloc debug/malloc hooks to handle this
new restrictive behavior.
Bug: 126944692
Test: Ran bionic unit tests.
Test: Ran bionic unit tests with malloc hooks enabled (no new tests fail).
Test: Ran bionic unit tests with malloc debug enabled (no new tests fail).
Test: Ran malloc debug unit tests.
Change-Id: I4d50785928815679c781ca729f998454d76b9192
This new option causes an abort after malloc debug detects an error.
This allows vendors to get process coredumps to analyze memory for
corruption.
Bug: 123009873
Test: New test cases added for unit tests and config tests.
Change-Id: I6b480af7f747d6a82f61e8bf3df204a5f7ba017f
There is a hang when enabling leak_track since the dumping of the
leak data can wind up doing an allocation.
Add new system unit test to make sure this doesn't happen again.
Bug: 111146059
Test: Test program that leaks does not hang forever.
Test: Unit tests pass.
Change-Id: Icf99be58ba5db98ee124a471b957a086045f5870
Bump the version from v1.1 to v1.2 and add a build fingerprint line.
Update the heap dump documentation to match the new format and reflect
what made it in P and what made it in Q.
Update the unit tests for this change.
Add -O0 to unit test code to make it easier to debug.
Add an external function that can be used by the framework code
so that there is only one way to dump the heap.
Bug: 110095681
Test: Ran unit tests.
Test: Did a dump of a real process and verified fingerprint.
Test: Did a dump of a process without malloc debug enabled.
Change-Id: I769a476cbeaf4c85c5d75bd6d6385f0e3add948c
Merged-In: I769a476cbeaf4c85c5d75bd6d6385f0e3add948c
(cherry picked from commit c84a2a2601)
This adds a new option backtrace_full, when it is set, then it will use
libunwindstack.
Modify the dump to file data to dump the extra information from libunwindstack.
Along with the new dump file format, change the version to v1.1.
Updated document for new format of file data.
Add unit tests for the new functionality.
Bug: 74361929
Test: Ran unit tests.
Change-Id: I40fff795f5346bba7b9d7fde2e04f269ff4eb7f1
Changes
- Refactor the code so that only guards require creating a special header
for every pointer allocated.
- Store only a single copy of every backtrace. This saves memory so that
turning on the backtrace option doesn't result in 10X memory usage.
- Added new option track_allocs that only verifies pointers are valid for
free/malloc_usable_size/realloc.
- Remove suffix from test names.
- Add the TRACK_ALLOCS options to all guard options.
- Add new option verify_pointers that is a lightweight way to verify
pointers that are passed to allocation routines.
- Do auto-formatting of the code.
- Updated documentation for all of these changes.
Bug: 74361929
Test: Ran unit tests.
Test: Ran libmemunreachable unit tests.
Test: Ran an app with backtrace enabled.
Change-Id: I3246c48ae4f9811f64622d90d0a9b4d9d818702c
While this was never an inline, this function alone has caused most of
the bug reports related to _FILE_OFFSET_BITS=64. Providing an inline
for it should allow a lot more code to build with _FILE_OFFSET_BITS=64
when targeting pre-L.
Test: make checkbuild
Test: built trivial cc_binary for LP32 against API 14 with
_FILE_OFFSET_BITS=64 set
Bug: lots
Change-Id: I8479d34af4da358c11423bee43d45b59e9d4143e
For non-zygote spawned processes, we might want to dump the backtrace
data. Provide a method to send a signal to a process and then dump the
data to a file.
Adds a method to dump the backtrace data on exit.
Update documentation and explain format of heap dump data.
Test: Ran unit tests, enabled new options and used them.
Change-Id: Ie2fa706694160731afe02c1382b037d06df1d069
Add the mallopt function, and only a single option so far.
Bug: 36401135
Test: Built and booted bullhead.
Test: Ran jemalloc unit tests.
Test: Ran bionic unit tests.
Test: Ran a test that allocated and free'd a large piece of memory,
Test: and verified that after changing the parameter, the PSS
Test: sticks around (decay timer set to 1), the PSS is purged (decay
Test: timer set to 0).
Change-Id: I6927929b0c539c1023d34772d9e26bb6a8a45877
This should make it easier to add new options, and to add options that
are complex. For example, I want to modify the behavior of
record_allocs_file so that it also enables record_allocs to a default
state.
Test: All unit tests pass.
Test: Enable the backtrace option and restart.
Change-Id: Idf5cdeed06ade3bc2c8ae39d228734bf65209b4f
Previously malloc debug can be enabled only using global settings
accessible to the root user only. This CL adds a new option to enable
it using environment variables making it possible to use it with pure
native (shell) applications on production builds (from shell user) and
prepares it for using it from logwrapper on production devices.
Remove the old environment variable and property since they are not
necessary.
Test: Enable malloc debug using environment variable and verify
Test: that it only affects the commands launched from the shell.
Test: Enable malloc debug using the property variable and verify
Test: that it affects all commands.
Test: Run all unit tests in 32 bit and 64 bit.
Change-Id: Iecb75a3471552f619f196ad550c5f41fcd9ce8e5
This option adds the ability to record all of the allocation requests
and dump them to a file when a signal is sent to the process.
Included in this change, redo the option processing to add a new
string option.
Bug: 27747898
Change-Id: Ida043362e38b5eb1d459c99db9c2581015dab366
- Move all ScopedDisableDebugCalls into the debug_XXX calls. This avoids
any issues that might arise where every part of the code needs to properly
guard anything that might allocate. Instead everything is already guarded.
- Add a pointer to debug_data in all of the XXData classes. This avoids
calling individual functions passing in the debug_data pointer.
- Flip the NO_HEADER_OPTIONS to an explicit HEADER_OPTIONS list since fewer
options actually require a header.
- Move the extern of g_debug to the DebugData.h header.
Change-Id: Ia213a391b4a44d9ce122a709d09fe4f1b5426f36
The free track mechanism could fail if, at the same time a free occurs,
another thread is trying to free and verify the same allocation. This
doesn't work if the freed allocation is added to the list and we still
do work on it. The fix is to only add to the free list when we are done
with the allocation.
Also fix a problem where the usable size is computed incorrectly because
two of the arguments where reversed.
In addition, add a check that the allocation being verified has the correct
tag before trying to check the body of the allocation.
Add a test to catch the original failure.
Add a test for the tag being different.
Bug: 27601650
Change-Id: Ie9200677d066255b8e668a48422f23f909f4ddee
The over flow check for nmemb * bytes in debug_calloc is incorrect,
use the builtin overflow functions to check for multiplication and
addition overflow.
Change-Id: I3f1c13102621bc5380be1f69caa88dba2118f3cb
(cherry picked from commit 239838608d)
0 byte allocations can cause problems if they are immediately followed
by another allocation with no header, as both allocations will have the
same address. Treat 0 byte allocations as 1 byte allocations so that
debug_iterate will return separate addresses for them.
Bug: 27578580
Change-Id: Ia8dc3481fa7062391e9b3ae58a36e8d47e7ee557
(cherry picked from commit 15af478080)
malloc_iterate with malloc debug enabled was returning allocation sizes
using the requested size instead of usable size. If anything wrote
pointers above the requested size but below the usable size, those
pointers would be invisible to libmemunreachable and referenced
allocations could be reported as a leak.
Bug: 27107100
Change-Id: I6b19fd631f68ce93b3aee408cc3d296ec457bd01
(cherry picked from commit b23c6606d2)
Add backtrace_string to convert a malloc_debug backtrace to a string.
Also move the backtrace functions to libc_malloc_debug_backtrace so that
libmemunreachable can reuse them.
Change-Id: I5ad67001c0b4d184903c762863a8588181d4873b
In order to enforce this constraint:
The pointer returned if the allocation succeeds shall be suitably
aligned so that it may be assigned to a pointer to any type of object
and then used to access such an object in the space allocated.
Force all allocations on 32 bit systems to have 8 byte alignment,
and all allocations on 64 bit systems to have 16 byte alignment.
Add a test to verify that the allocator returns the correct alignments.
Bug: 26739265
Change-Id: I9af53279617408676b94e4ec6481b3ed7ffafc6a
Add a pthread_atfork handler to malloc_debug to lock the TrackData mutex
during fork and reset it in the child. Ensures that the TrackData is
consistent when forking from a multi-threaded process, and that the
mutex is in a defined state in the child.
Change-Id: I0dfed30045a28551217ceac227d33b2e18894932
Included in this change:
- Change the tag when a pointer is freed so it's easy to detect if
an already freed pointer is being used.
- Move the free backtrace out of the header. This backtrace is only
used under only some circumstances, so no need to allocate space
in all headers for it.
- Add new option free_track_backtrace_num_frames to specify how many
frames to record when the free occurs. This removes the dependency
on the backtrace option to get backtraces.
Bug: 26739265
Change-Id: I76f5209507dcf46af67ada162a7cb2bf282116f2
The major components of the rewrite:
- Completely remove the qemu shared library code. Nobody was using it
and it appears to have broken at some point.
- Adds the ability to enable/disable different options independently.
- Adds a new option that can enable the backtrace on alloc/free when
a process gets a specific signal.
- Adds a new way to enable malloc debug. If a special property is
set, and the process has an environment variable set, then debug
malloc will be enabled. This allows something that might be
a derivative of app_process to be started with an environment variable
being enabled.
- get_malloc_leak_info() used to return one element for each pointer that
had the exact same backtrace. The new version returns information for
every one of the pointers with same backtrace. It turns out ddms already
automatically coalesces these, so the old method simply hid the fact
that there where multiple pointers with the same amount of backtrace.
- Moved all of the malloc debug specific code into the library.
Nothing related to the malloc debug data structures remains in libc.
- Removed the calls to the debug malloc cleanup routine. Instead, I
added an atexit call with the debug malloc cleanup routine. This gets
around most problems related to the timing of doing the cleanup.
The new properties and environment variables:
libc.debug.malloc.options
Set by option name (such as "backtrace"). Setting this to a bad value
will cause a usage statement to be printed to the log.
libc.debug.malloc.program
Same as before. If this is set, then only the program named will
be launched with malloc debug enabled. This is not a complete match,
but if any part of the property is in the program name, malloc debug is
enabled.
libc.debug.malloc.env_enabled
If set, then malloc debug is only enabled if the running process has the
environment variable LIBC_DEBUG_MALLOC_ENABLE set.
Bug: 19145921
Change-Id: I7b0e58cc85cc6d4118173fe1f8627a391b64c0d7
