| 7c170bb1b1abf5a055ad8137e8531b5bcd0eb5ce |
|
05-May-2018 |
Martijn Coenen <maco@google.com> |
Create temp refs on proxies.
Fixes the following scenario:
1) Thread T1 receives a binder proxy and adds a local strong/weak ref
2) Thread T1 queues BC_ACQUIRE/BC_INCREFS (but does not flush)
3) Thread T1 hands the binder proxy off to another thread T2, which
adds another strong ref
4) Thread T1 now drops its own ref to the proxy, but it doesn't get
destructed because T2 has its own ref to it
6) Thread T2 runs and drops its own ref to the proxy, which causes a
BC_RELEASE/BC_DECREFS to be queued to the driver.
Now, if T1 writes its command queue to the driver first, everything
is fine, because a BC_ACQUIRE/BC_INCREFS will be followed by a
BC_RELEASE/BC_DECREFS. However if T2 writes its command queue first,
BC_RELEASE/BC_DECREFS will be sent first, and the driver will delete
the reference prematurely.
Fix this by temporarily holding a weak/strong ref until the initial
BC_ACUIRE/BC_INCREFS is flushed to the driver.
Bug: 78437964
Test: sailfish builds, boots, no invalid ref messages on dumpsys
Change-Id: Ica5efaf9fa2785423c37abb1456575d97f1cd85f
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| ea0090a0392db2f3b7b047647d48b01885722cda |
|
02-Nov-2017 |
Martijn Coenen <maco@google.com> |
Revert "binder: send BC_REPLY and BC_FREE_BUFFER together"
This reverts commit 0494d6ab6f4b85eb28eeb71737398ba0b862a822.
Change-Id: I115dd7cd3e11bb573e17b3859133e0cad1afef49
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 0494d6ab6f4b85eb28eeb71737398ba0b862a822 |
|
02-Nov-2017 |
Ganesh Mahendran <opensource.ganesh@gmail.com> |
binder: send BC_REPLY and BC_FREE_BUFFER together
In current BR_TRANSACTION handling logic, BC_REPLY and BC_FREE_BUFFER
are sent to kernel seperately which may introduce latency for freeing
buffer.
I think it's better to free buffer asap, this change fixes this
by putting BC_REPLY and BC_FREE_BUFFER in the same buffer. And then
call waitForResponse() to send them to kernel together.
After this, function sendReply() will not be called by anyone. So
sendReply() is also removed.
Below is the test result of "binderThroughputTest -w 100"
Env: android 7.1.2, 2G ram
---------------------
---> base:
iterations per sec: 31917.7
collecting results
average:2.83375ms worst:572.205ms best:0.054219ms
50%: 2.14844 90%: 5.27344 95%: 6.83594 99%: 12.3047
iterations per sec: 32142.3
collecting results
average:2.8186ms worst:550.884ms best:0.054948ms
50%: 2.53906 90%: 5.27344 95%: 6.83594 99%: 11.1328
iterations per sec: 31704.5
collecting results
average:2.85129ms worst:600.984ms best:0.053854ms
50%: 2.53906 90%: 5.27344 95%: 6.83594 99%: 12.3047
iterations per sec: 31932.3
collecting results
average:2.82549ms worst:563.098ms best:0.052396ms
50%: 2.53906 90%: 5.66406 95%: 6.83594 99%: 11.1328
---> patched:
iterations per sec: 32062.5
collecting results
average:2.85475ms worst:599.177ms best:0.054895ms
50%: 2.14844 90%: 5.27344 95%: 6.44531 99%: 12.6953
iterations per sec: 32241.7
collecting results
average:2.82851ms worst:634.984ms best:0.05651ms
50%: 2.53906 90%: 5.66406 95%: 6.83594 99%: 10.7422
iterations per sec: 31939
collecting results
average:2.86589ms worst:628.861ms best:0.058645ms
50%: 2.53906 90%: 5.66406 95%: 6.83594 99%: 11.1328
iterations per sec: 32149.4
collecting results
average:2.80785ms worst:658.889ms best:0.055573ms
50%: 2.53906 90%: 5.66406 95%: 6.83594 99%: 10.7422
It seems patched is better.
Change-Id: I0ef88864f4dbc63f3cdd2eba345b5dce6140b2d3
Suggested-by: Martijn Coenen <maco@google.com>
Signed-off-by: Ganesh Mahendran <opensource.ganesh@gmail.com>
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 58e5daaed8b446bdcf937a5eb368d6623d33f423 |
|
11-Oct-2017 |
Ganesh Mahendran <opensource.ganesh@gmail.com> |
binder: remove unnecessary err check
In function IPCThreadState::transact(), data.errorCheck() will
be executed twice. Since IPCThreadState::transact() is the
critical path for binder call, it is better to do data.errorCheck()
once.
This patch removes the first check at the beginning of
IPCThreadState::transact(), the effect of this change is that LOG_ONEWAY(...)
will be executed in error case("data.errorCheck() != NO_ERROR")
I think this is not a problem. As "data.errorCheck() == NO_ERROR" is the
normal case(which will execute LOG_ONEWAY()), and even in error case,
we print a log about src/dst pid is not a bad idea.
Change-Id: I7b892a2294774c55ce0df56edee6a820f82c6f13
Signed-off-by: Ganesh Mahendran <opensource.ganesh@gmail.com>
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 9cec85be1921a74e19cadd8b9b5a563eec6f82a5 |
|
20-Sep-2017 |
Manoj Gupta <manojgupta@google.com> |
Silence a false positive memory leak warning.
This silences a false positive memory leak warning:
frameworks/native/libs/binder/IPCThreadState.cpp:689:5: warning: Potential
memory leak [clang-analyzer-cplusplus.NewDeleteLeaks]
Bug: b/27101951
Test: The warning is gone.
Change-Id: I2d7bbb943521df86debb3a0e1b4b7457fa847098
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 3c4f8434fa5f719e465c88b23a0cc6cb7697e0ea |
|
15-Aug-2017 |
Martijn Coenen <maco@google.com> |
Merge "Don't lose BR_RELEASE/BR_DECREFS commands."
am: 372bccede8
Change-Id: Ib3c41153b1c5640f9b9e938c1b4872d2fa441e9c
|
| 0791fbf85aef8f40cee0821223c46afddd3fa464 |
|
08-Aug-2017 |
Martijn Coenen <maco@google.com> |
Don't lose BR_RELEASE/BR_DECREFS commands.
BR_RELEASE/BR_DECREFS commands are stored in mPendingStrongDerefs
and mPendingWeakDerefs, respectively. During processPendingDerefs(),
we actually execute the corresponding decStrong()/decWeak() operations.
The problem is that when we're done, we clear() the mPending vectors
without checking if new entries have been added to them. This can
happen, because decStrong()/decWeak() might cause destructors to run,
which in turn can invoke outgoing transactions that may result in
a BR_RELEASE/BR_DECREFS being queued.
Bug: 63079216
Test: binderLibTest
Change-Id: Ib1deca3f317f8b5068b4b9eddfc4219b9ec87740
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 2b6317436d2035ce98331906aaaca87e6026c9c8 |
|
05-May-2017 |
Martijn Coenen <maco@google.com> |
Correctly set minimum priority on nodes.
It looks like libbinder has always set the minimum priority
for a node to 0x7f (127). This is an invalid value as far
as the kernel is concerned, since the current driver only
accepts positive nice values (0..19). The effect of using
0xf7 is that the nice value was clamped to MAX_NICE (19) -
so effectively the kernel never raised the priority
based on the min_priority setting of a node. Correct this
by just using MAX_NICE directly.
Additionally, a recent change removed the use of the
gDisableBackgroundScheduling flag, which system_server
used to make sure that incoming transactions were not
executed at a lower priority. That behavior is actually
still desired, so this change sets the min_priorirty
value to 0 for processes that have requested background
scheduling to be disabled. The effect of that is that
all transactions into those nodes will be executed with
a priority of at least nice 0 (and potentially higher,
depending on the priority of the caller).
Bug: 37677242
Test: verified /d/binder output
Change-Id: I7cbfd309d04cbd052f868fbfe0930529ff21a48a
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| d429f4a2d5cf2db093bb0e35cb651d764d1973c8 |
|
07-Mar-2017 |
Tim Murray <timmurray@google.com> |
binder: remove cgroup interactions
Having binder change cgroups on the fly is too expensive in the common case to be useful.
Test: builds
bug 34825105
Change-Id: Ia692c9d2ab261d852cb2fac8f27d6779d0e2b057
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| a43ee156098776698a39882bb00ecebc555948d0 |
|
25-Oct-2016 |
liangweikang <liangweikang900911@gmail.com> |
binder: print the bad cmd error to android log
https://code.google.com/p/android/issues/detail?id=226068
Signed-off-by: liangweikang <liangweikang900911@gmail.com>
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 73a7dde7c3e5c8437f396deeb200f250b3346be7 |
|
18-Oct-2016 |
songjinshi <songjinshi@xiaomi.com> |
[binder]: Correct the index for getReturnString
The cmd code can not as the index drectly.
Change-Id: Icddad46faef7a77acbd7edf118c081ac10501f47
Signed-off-by: songjinshi <songjinshi@xiaomi.com>
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| c734143ca5bf7acfea782b79cfd6700af206a9ca |
|
19-Oct-2016 |
Wei Wang <wvw@google.com> |
Fix log function for potential overflow
On LP64 system, unsigned long can be 64bit
Also clean out unused static variables.
Test: flash on ARM64 device
Bug: 32181382
Change-Id: I44b7ea8a6588c475a3979d7bddeb08da7f54c27a
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 249f373defa993eb9c8358e4d2e0b620f2c10d57 |
|
16-Apr-2016 |
Colin Cross <ccross@android.com> |
Merge "Log when binder thread pool is starved" into nyc-dev
am: 43dcac7
* commit '43dcac7500833c465c417805adc4693c60c3bb5e':
Log when binder thread pool is starved
Change-Id: I420d31273aec69eab9d8e4faa10001ef48b17aa8
|
| 96e8322f4d11af8538d6d0db5f1a11338b8aee9d |
|
15-Apr-2016 |
Colin Cross <ccross@android.com> |
Log when binder thread pool is starved
Log when the number of threads executing binder commands is equal to the
maximium size of the thread pool for >100ms.
Bug: 28201939
Change-Id: I892863d8a81c06e362d4ae18ab08485fdec3c0bb
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 7f0c6d6e19565e512fc42c2371bc99f4c5e3fe70 |
|
21-Mar-2016 |
Dianne Hackborn <hackbod@google.com> |
Merge "Fix issue #27252896: Security Vulnerability -- weak binder" into nyc-dev
am: b49576d
* commit 'b49576d5241859a3cf03c9f34b8030ad25d3806b':
Fix issue #27252896: Security Vulnerability -- weak binder
|
| c11146106f94e07016e8e26e4f8628f9a0c73199 |
|
21-Mar-2016 |
Dianne Hackborn <hackbod@google.com> |
Fix issue #27252896: Security Vulnerability -- weak binder
Sending transaction to freed BBinder through weak handle
can cause use of a (mostly) freed object. We need to try to
safely promote to a strong reference first.
Change-Id: Ic9c6940fa824980472e94ed2dfeca52a6b0fd342
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| ff405785386ed8bdee50c4afdc4a4f9a73bcb81e |
|
09-Mar-2016 |
zhongjie <zhongjie.shi@intel.com> |
Release resources when repeating dlopen-dlcose.
When the libbinder.so is dlopen and dlclose many
times, we need to delete the pthread_key and
munmap the memory mapping.
Otherwise, it will crash as the leakage causes
the pthread_key or memory mapping surpass their
limit.
Change-Id: Ice64d4ef685c1f665350286f9e5ddf790beac236
Signed-off-by: zhongjie <zhongjie.shi@intel.com>
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| f31a3eb3290c4154bd04113c8bb859a499f288b6 |
|
01-Feb-2016 |
Andreas Gampe <agampe@google.com> |
Binder: Add log output in IPCThreadState::self()
In the failure case, and on shutdown, add log output. It's highly
likely the client will crash when a null value is returned. Displaying
a possible root cause helps diagnosing issues.
Bug: 26865304
Change-Id: Ie5255ac50d6dcdf9f918dcef785788ea795a4791
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 30dddd6f49b04ac5160c1de702cafe09b42ce1d1 |
|
13-Aug-2015 |
Elliott Hughes <enh@google.com> |
am 6a846dd5: am 5f5a0e0f: Merge "Lose HAVE_ANDROID_OS from frameworks/native."
* commit '6a846dd51885df861f4e1e30fcca86b9dd2f4092':
Lose HAVE_ANDROID_OS from frameworks/native.
|
| 6071da7ef84c60645572654504813d492b8b21d5 |
|
13-Aug-2015 |
Elliott Hughes <enh@google.com> |
Lose HAVE_ANDROID_OS from frameworks/native.
Change-Id: I8e6af2a46a9d875192fc7957ada9b5f66d84bad6
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| a3206e690bcfe5c982a0d7fff7c845b83a3ed91f |
|
21-Apr-2015 |
Wale Ogunwale <ogunwale@google.com> |
Fixed build error with clang/llvm.
Change-Id: Ideacdc974ebad542df724464ccba9fcfb2b7ea91
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 376b822728c582a9deb26ce339deb5204ae4d22e |
|
14-Apr-2015 |
Wale Ogunwale <ogunwale@google.com> |
Added IPCThreadState::blockUntilThreadAvailable() method.
Will be used by the system_server watchdog to monitor the
availability of binder threads in the process to handle
incoming IPC requests.
Bug: 19297165
Change-Id: I39175f3869ad14da5620fddb47f454e6e4ee2b25
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 8fb2d25ff3e9ade7c03ebe52a89c671ee75ce484 |
|
27-Jan-2015 |
Yabin Cui <yabinc@google.com> |
Kill HAVE_PTHREADS.
Bug: 19083585
Change-Id: I355491de945590f43c82bdcb7968b01b4bff6e06
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 9c634fd2d13722106d475b000b05babcce1f65be |
|
26-Nov-2014 |
Dan Stoza <stoza@google.com> |
binder: Return pid_t/uid_t instead of int
Makes IPCThreadState::getCalling{Pid,Uid} return a more
appropriate type.
Change-Id: I97f8a83c1c0722bc1afbf8a6df1a288f8f1f0d2c
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 74debb0b8fc47d443aaf7fe79f7b67f53f735f7f |
|
25-Nov-2014 |
Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org> |
Fix build with clang in C++11 mode
Now that we're in C++11 mode by default, clang complains about switch
statements with case values not matching the type -- since some
binder_driver_return_protocol values are > 0x7fffffff, we need to make
the switch statements operate on uint32_t rather than int32_t.
BUG: 18466763
Change-Id: Iedbfd5c7a3d3d9f087d2eab4ff21343ad7a2a448
Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 0b53b5c0b34ff0e14e05e7a4613c7364136f47a5 |
|
24-Nov-2014 |
Andreas Gampe <agampe@google.com> |
Binder: Remove unused functions & variables
For build-system CFLAGS clean-up, remove unused functions and
variables.
Change-Id: I88abc2bc99e6b91c7aa56470ff9671bf8bf29f8d
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| ac5cd96fe24690d9e0ae823c77334b8579139aa2 |
|
04-Nov-2014 |
Chih-Hung Hsieh <chh@google.com> |
am dbab409c: Merge "Avoid two warnings from clang."
* commit 'dbab409c49ca5dd6686912d0b3eeb19ef1c37a3e':
Avoid two warnings from clang.
|
| 8e5337de2c67c24f1e5c6144512e2a9ba9fefe12 |
|
24-Oct-2014 |
Chih-Hung Hsieh <chh@google.com> |
Avoid two warnings from clang.
Clang compiler warns about undefined static functions like getReturnString
before all references are removed inside IF_LOG_COMMANDS.
To remove such warnings, we can leave those static functions defined
and the compiler will remove them at the end when they are not used.
Add a cast to long before converting from int to void*,
to avoid type cast warnings on 64 bit targets.
BUG: 17043248
Change-Id: Ica31101522279d2c8eacc90e2939767002fb827b
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| e4499526eee52a6584aba9382783e9fba327a9bc |
|
30-Sep-2014 |
Dianne Hackborn <hackbod@google.com> |
am 5ee2c9d2: Work on issue #17656716: Unhandled exception in Window Manager
* commit '5ee2c9d20c0fbecf6b4a482eb8d8bfdf85d0d424':
Work on issue #17656716: Unhandled exception in Window Manager
|
| 5ee2c9d20c0fbecf6b4a482eb8d8bfdf85d0d424 |
|
30-Sep-2014 |
Dianne Hackborn <hackbod@google.com> |
Work on issue #17656716: Unhandled exception in Window Manager
Try to clean up the code paths coming in and out of binder IPCs to
plug any places where we could disrupt the gather flag of a thread,
causing it to keep gathering stack crawls (which is the thing that
is causing our strict mode data to become so large).
We now take care of saving and restoring this state in the core
IPC code path, not at the Java layer.
Change-Id: I73d564778da127bdce00f304225930e7f2318293
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| fa6ad07c4ed44244f5220a422819919bc37528ba |
|
18-Aug-2014 |
Elliott Hughes <enh@google.com> |
Use gettid directly in binder.
Bug: 17048545
Change-Id: I1df378135961e085a52f6cc5694ad74cd73aac19
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 317ff37cf79d31170fa7320f0ba5ded1ec043e4c |
|
29-May-2014 |
Mark Salyzyn <salyzyn@google.com> |
binder: rm utils/Debug.h
Change-Id: I16931919a26c16a0d701771c93fa7fe1c6f60ef8
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 07fd0f195db6d341cab4e54257f508d802c98832 |
|
19-Feb-2014 |
Arve Hjønnevåg <arve@android.com> |
Binder: Fix some valgrind errors.
When using 64 bit binder pointers, only initializing the 32 bit
handle, in a stack allocated struct, will pass uninitialized stack
data to the kernel and other processes.
Change-Id: I3432d9d36bb251d8ddb0a863661aeb80aabb3d92
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 87b30d0447829167b2d83f4f61f702638d937524 |
|
19-Feb-2014 |
Arve Hjønnevåg <arve@android.com> |
Binder: Don't cast directly from a pointer to binder_uintptr_t
When using the 64 bit binder interface from a 32 bit process the
pointer may get sign extended and cause the kernel to fail to read
from it.
Change-Id: I90fcf53880e2aa92e230a9723f9b3f7696170e32
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 11cfdccfd3cfceb08732909a1489419ff0229694 |
|
15-Feb-2014 |
Arve Hjønnevåg <arve@android.com> |
Binder: Disable attemptIncStrongHandle
The driver does not support BC_ATTEMPT_ACQUIRE and will return an error.
IPCThreadState does not handle driver errors, and will resend the failed
command blocking all other commands.
Change-Id: I643986037341821b27b62dc82df933844f4842b8
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 6f4f3ab36c5ed1df84eb3a9f7475f0ac42952f58 |
|
06-Feb-2014 |
Colin Cross <ccross@android.com> |
binder: fix all warnings
Fix warnings related to casting pointers to ints, using %d or %ld to
print size_t/ssize_t, and unused parameters.
Change-Id: I7a13ba83d402952989c1f795cd9e880a95b98d9e
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 84e625ac1e01f5a9c3ed16664da6498212ed662b |
|
29-Jan-2014 |
Arve Hjønnevåg <arve@android.com> |
Binder: Use 64 bit pointers in 32 processes if selected by the target
Uses new kernel header where void * has been replaced by binder_uintptr_t
Change-Id: Icfc67c2a279269f700343bd9246fd7cb94efe2c1
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| f683e0163a84d93448b9388126902242367cd961 |
|
05-Nov-2013 |
Serban Constantinescu <serban.constantinescu@arm.com> |
Binder: Make binder portable
Changes include
- Binder attempts to cast pointers to a int datatype
which is not sufficient on a 64-bit platform.
- This patch introduces new read/write functions into
Parcel that allow pointers to be written using the
uintptr_t datatype for compile-time data type size
selection.
- Change access specifier for the methods above.
- Binder uses the 64bit android_atomic_release_cas64
(aka cmpxchg)
Change-Id: I595280541e0ba1d19c94b2ca2127bf9d96efabf1
Signed-off-by: Matthew Leach <matthew.leach@arm.com>
Signed-off-by: Serban Constantinescu <serban.constantinescu@arm.com>
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| ada11c5e67cf8587ad3b53dd99c27455ef26291f |
|
13-Jul-2013 |
Jeff Brown <jeffbrown@google.com> |
Make getCallingUid/Pid const.
Change-Id: I1853b21eaa45d85274189dfd72f73fec48d0d6b7
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 8d96cab8bfc1b40a8e05c6f43b485750a5ae0a52 |
|
26-Jun-2013 |
Todd Poynor <toddpoynor@google.com> |
binder: add polling / single-threaded operation
This is currently safe to do only for processes that disallow any binder
threads to be created: setThreadPoolMaxThreadCount(0).
Change-Id: I8a27f3cf26f4d51edb7f222af487ac256cbcab65
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| fc1f308ca1d679f9e9823a848e6396a23513f2fb |
|
12-Jun-2013 |
Jeff Tinker <jtinker@google.com> |
am de836890: am ef07386e: Prevent IPCThreadState::joinThreadPool from spinning on unexpected error
* commit 'de836890abc7b5381e285833052375cbb85730cb':
Prevent IPCThreadState::joinThreadPool from spinning on unexpected error
|
| ef07386e2fca73680214ececc3c9c0ecbb0f6d88 |
|
11-Jun-2013 |
Jeff Tinker <jtinker@google.com> |
Prevent IPCThreadState::joinThreadPool from spinning on unexpected error
Adds logging to help determine what is happening to the /dev/binder
fd and dump the process state when it happens.
bug: 8912673
Change-Id: I2aa0c66fc499e91e0bf9ee4ae20404bec35adc82
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 002e1e58dfe19dd3e49a59c6827cbf51573941a2 |
|
07-May-2013 |
Mathias Agopian <mathias@google.com> |
libutils clean-up
Change-Id: I6ff4cfc736751de2912c697f954e45e275f2d386
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 0bed1f541d5a8284691d2296754604cde0723787 |
|
07-Oct-2012 |
Dave Burke <daveburke@google.com> |
Revert "ugly, temporary, workaroung for a problem where a binder thread spins forever"
This reverts commit 0845d0245e09548110cacb0f20e9934753388aab
Change-Id: I395037cb9427cd11f7de6bb78fbdfa917fc6263a
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 0845d0245e09548110cacb0f20e9934753388aab |
|
06-Oct-2012 |
Mathias Agopian <mathias@google.com> |
ugly, temporary, workaroung for a problem where a binder thread spins forever
Bug: 7289992
Change-Id: I0c3d482a1af57e5f444be2ba7f2751ac3e954af2
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 8210185fe337ca9d5b01f2bff8590ea60984c31e |
|
31-Aug-2012 |
Dianne Hackborn <hackbod@google.com> |
No longer need "original calling uid".
Change-Id: Ifc4a89dd088609a5a8553f6ac6553174e09e8222
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| db1597a98958b78dc0d8eced19ae1406382b70d6 |
|
17-Feb-2011 |
Johannes Carlsson <johannes.carlsson.x@sonyericsson.com> |
Fix shutdown sequence to avoid SIGSEGV when running am command
When the app_process is shutting down the main thread will close the
binder fd while pool threads are executing an ioctl (in
IPCThreadState::stopProcess called by AppRuntime::onStarted in
app_main.c).
The binder driver will then return all pending calls in ioctl
without any error and with a command. One of the threads gets a
BR_SPAWN_LOOPER which will create a new thread (the other thread
gets a BR_NOOP). This new thread then calls
vm->AttachCurrentThread. Usually this results in a log entry with
"AndroidRuntime: NOTE: attach of thread 'Binder Thread #3' failed",
but sometimes it also causes a SIGSEGV. This depends on the timing
between the new thread an the main thread that calls DestroyJavaVM
(in AndroidRuntime::start).
If IPCThreadState.cpp is compiled with "#define LOG_NDEBUG 0" the
pool thread will loop and hit the
ALOG_ASSERT(mProcess->mDriverFD >= 0) in
IPCThreadState::talkWithDriver.
Crashes like this has been seen when running the am command and
other commands that use the app_process.
This fix makes sure that any command that is received when the driver
fd is closed are ignored and IPCThreadState::talkWithDriver instead
returns an error which will cause the pool thread to exit and detach
itself from the vm. A check to avoid calling ioctl to a fd with -1
was also added in IPCThreadState::threadDestructor.
Another solution might be to change the binder driver so that it
returns an error when the fd is closed (or atleast not a
BR_SPAWN_LOOPER command). It might also be possible to call exit(0)
which is done when System.exit(0) is called from java.
Change-Id: I3d1f0ff64896c44be2a5994b3a90f7a06d27f429
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| a26e1cfbbcb7dc5e0b1cd79a815362d6a048ef26 |
|
16-Mar-2012 |
Glenn Kasten <gkasten@google.com> |
Scheduling group cleanup
Remove C++ APIs androidSetThreadSchedulingGroup and
androidGetThreadSchedulingGroup, and the ANDROID_TGROUP_* constants.
Former callers of these should now use the C APIs set_sched_policy and
get_sched_policy, and the SP_* constants.
Note: debug.sys.noschedgroups is not supported by the C APIs,
this needs to be discussed.
Change-Id: I32bbfc539ef4090faf9ef0320380e8cca9eae07c
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 4bcb57502c5a7da53ef84228aceeffb54d36e06b |
|
06-Mar-2012 |
Glenn Kasten <gkasten@google.com> |
Remove obsolete references to sched_policy.h
As part of scheduling policy cleanup, remove or isolate
all references to the scheduling policy APIs.
Change-Id: Ia1ea2fe711a399039f25217309e061267744b856
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 4e975bb488bb3947703e95cb33a4838adda680f7 |
|
04-May-2011 |
Amith Yamasani <yamasani@google.com> |
Multi-user - 1st major checkin
Switching activity stacks
Cache ContentProvider per user
Long-press power to switch users (on phone)
Added ServiceMap for separating services by user
Launch PendingIntents on the correct user's uid
Fix task switching from Recents list
AppWidgetService is mostly working.
Commands added to pm and am to allow creating and switching profiles.
Change-Id: I15810e8cfbe50a04bd3323a7ef5a8ff4230870ed
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 6726347e8950d34ae162fb8d6a3680a871d359e2 |
|
09-Jan-2012 |
Steve Block <steveblock@google.com> |
Rename LOG_ASSERT to ALOG_ASSERT DO NOT MERGE
See https://android-git.corp.google.com/g/157519
Bug: 5449033
Change-Id: I8ceb2dba1b031a0fd68d15d146960d9ced62bbf3
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| a19954ab377b46dbcb9cbe8a6ab6d458f2e32bca |
|
04-Jan-2012 |
Steve Block <steveblock@google.com> |
Rename (IF_)LOGI(_IF) to (IF_)ALOGI(_IF) DO NOT MERGE
See https://android-git.corp.google.com/g/156801
Bug: 5449033
Change-Id: Ib08fe86d23db91ee153e9f91a99a35c42b9208ea
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| a63ee4c2fdc4869fd1169d22e79ae9f1220c5bdb |
|
02-Dec-2011 |
Ben Cheng <bccheng@google.com> |
am f1a4c48b: am e21dabf6: Merge "Initialize bwr.read_buffer in IPCThreadState::talkWithDriver." into ics-mr1
* commit 'f1a4c48b886d53e2afceb0620215e3792af79fd5':
Initialize bwr.read_buffer in IPCThreadState::talkWithDriver.
|
| d640f89205e17d9f19b11c2954862f0945a40e7b |
|
02-Dec-2011 |
Ben Cheng <bccheng@google.com> |
Initialize bwr.read_buffer in IPCThreadState::talkWithDriver.
I/valgrind( 1309): ==1310== Syscall param
ioctl(BINDER_WRITE_READ).read_buffer points to uninitialised byte(s)
I/valgrind( 1309): ==1310== at 0x480E670: __ioctl (__ioctl.S:10)
I/valgrind( 1309): ==1310== Address 0xbda651b4 is on thread 1's stack
Change-Id: I02893df7b5786b6b2dbd9659f5706d7171295ab2
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 9f760150f6e0f39b9923cfdc875373606839ee00 |
|
12-Oct-2011 |
Steve Block <steveblock@google.com> |
Rename (IF_)LOG() to (IF_)ALOG() DO NOT MERGE
See https://android-git.corp.google.com/g/#/c/141576
Bug: 5449033
Change-Id: I42575e7c29cf1c0f465c357a5c97ab118df6f473
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| aefc9cda45adb00ec1cb2c8fbd03c50fd48499fb |
|
31-Aug-2011 |
Andy McFadden <fadden@android.com> |
Resurrect verbose binder logging
Updated the command name lists, and masked off the additional bits in
the command word when doing the name lookup.
Made descriptor values easier to grep for and consistent with kernel
output (i.e. decimal rather than hex). Attempt to show transaction
descriptors as such (they're in a union with a pointer).
Also, the writeLines() function in Static was using a no-op
logging call to write an iovec. It looks like all callers are using
N=1, so I just added a log for the first string.
Bug 5155269
Change-Id: I417b8d77da3eb6ee1d2069ba94047210f75738bc
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| d547432f98889d57a8c37e91090664efe63da436 |
|
21-Apr-2011 |
Evgeniy Stepanov <eugenis@google.com> |
Make sure binder ioctl structs don't contain uninitialized values.
Change-Id: I8a678f91262417bb120e65e32c244ce1512b46c2
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 1b6084354710ca52c04ea361d751d19174802d57 |
|
14-Dec-2010 |
Brad Fitzpatrick <bradfitz@android.com> |
Framework-side support for Dalvik "isSensitiveThread" hook.
Used in lock contention stats.
Bug: 3226270
Change-Id: Ie6f58d130a29079a59bdefad40b80304d9bc3623
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 67f78c4fe8f89b52199f509f36e42df0b1cdfe2d |
|
24-Sep-2010 |
Dianne Hackborn <hackbod@google.com> |
Some debugging support.
- New feature to "am monitor" to have it automatically launch
gdbserv for you when a crash/ANR happens, and tell you how to
run the client.
- Update dumpstate to match new location of binder debug logs
- Various commented out logs that are being used to track down
issues.
Change-Id: Ia5dd0cd2df983a1fc6be697642a4590aa02a26a5
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 5273603e98d2db3bac656b7bcf5352c04c86d62f |
|
31-Aug-2010 |
Brad Fitzpatrick <bradfitz@android.com> |
Don't propagate StrictMode over one-way Binder calls.
This was causing stack stitching problems where a one-way call with
violations followed by a two-way call without violations was getting
the previous one-way call's violation stack stitched on to the second
caller's stack.
The solution is a little more indirect than I would've liked
(preserving the binder's onTransact flags until enforceInterface) but
was seemingly necessary to work without changing the AIDL compiler.
It should also be sufficiently cheap, since no new calls to
thread-local IPCThreadState lookups were required. The additional
work is just same-thread getter/setters on the existing
IPCThreadState.
Change-Id: I4b6db1d445c56e868e6d0d7be3ba6849f4ef23ae
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| a877cd85b5a026384542e3271fc310d6a8fe24c6 |
|
08-Jul-2010 |
Brad Fitzpatrick <bradfitz@android.com> |
More StrictMode work, keeping Binder & BlockGuard's thread-locals in-sync.
Change-Id: Ia67cabcc17a73a0f15907ffea683d06bc41b90e5
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 702ea9d42f52fc145090c0f0bfbe64993e4b8b33 |
|
18-Jun-2010 |
Brad Fitzpatrick <bradfitz@android.com> |
Start of work on passing around StrictMode policy over Binder calls.
This is (intendend to be) a no-op change.
At this stage, Binder RPCs just have an additional uint32 passed around
in the header, right before the interface name. But nothing is actually
done with them yet. That value should right now always be 0.
This now boots and seems to work.
Change-Id: I135b7c84f07575e6b9717fef2424d301a450df7b
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 440fd870b20b4720e11c9ea71d2a3284f8cf2f59 |
|
19-Mar-2010 |
Christopher Tate <ctate@google.com> |
Ensure that binder incalls to the system process keep the fg cgroup
On binder incalls, the handler thread is given the caller's priority by the
driver, but not the caller's cgroup. We have explicit code that sets the
handler's cgroup to match the caller's, *except* that the system process
explicitly disables this behavior. This led to a siuation in which we were
running binder incalls to the system process at nice=10 but cgroup=fg.
That's fine as far as it goes, except that if a GC happened in the handler
thread, it would be promoted to foreground priority and cgroup both, to avoid
having the GC take forever. Then, when GC finished, the original priority
is reset, and the cgroup set *based on that priority*. This would push the
handler thread into nice=10 cgroup=bg_non_interactive -- which matches the
caller, but is supposed to be impossible in the system process.
The end result of this was that we could be running "lengthy" operations in
the system process in the background. Unfortunately, some of the operations
that wound up like this would hold important global system locks for up to
twenty seconds as a result, making the entire device unresponsive to input
for that period.
This CL fixes the binder incall setup to ensure that within the system process,
a binder incall is always begun from the normal foreground priority as well
as cgroup. In practice now the device still becomes laggy/sluggish when the
offending lock-holding time-consuming incall occurs, but since it still runs
as a foreground task it is able to proceed to completion within a short time
rather than taking 20 seconds.
Fixes bug #2403717
Change-Id: Id046aeabd0e80c48eef94accc37842835eab308d
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 8c6cedc9bc9a4b69616a79a95449f6f6b08c7bf1 |
|
08-Dec-2009 |
Dianne Hackborn <hackbod@google.com> |
Propagate background scheduling class across processes.
This is a very simply implementation: upon receiving an IPC, if the handling
thread is at a background priority (the driver will have taken care of
propagating this from the calling thread), then stick it in to the background
scheduling group. Plus an API to turn this off for the process, which is
used by the system process.
This also pulls some of the code for managing scheduling classes out of
the Process JNI wrappers and in to some convenience methods in thread.h.
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 07d69893e1677bd59a3461a0c4fcd3541563144d |
|
08-Nov-2009 |
Christopher Tate <ctate@android.com> |
Reset binder service threads' cgroup/priority after command completion
To prevent buggy command implementations from poisoning binder threads'
scheduling class & priority for future command execution, we now reset the
cgroup and thread priority to foreground/normal when a binder service thread
finishes executing the designated command.
Change-Id: Ibc0ab2485751453f6dc96fdb4eb877fd02796e3f
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 6dfe8f1ffad455363fff4fe69ced6da386970fbc |
|
06-Nov-2009 |
Evan Millar <emillar@google.com> |
Revert jparks code from IPCThreadState.
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| b5c41353337b3db2dc0ca0df163d5294af8281bc |
|
04-Nov-2009 |
Jason Parks <jparks@google.com> |
When a thread is about to be put back onto the thread pool ensure that it is in the foreground cgroup.
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| dcd3958c5086f757dc09472700ae7384efea7fc8 |
|
03-Nov-2009 |
Jason Parks <jparks@google.com> |
Add a warning when we leave threads in the binder thread pool in the background scheduling group.
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| d43b194b69fca6c81023effc921fcc7576bad496 |
|
17-Jul-2009 |
Marco Nelissen <marcone@google.com> |
Instead of using -1 for pid and uid in the simulator, and then having
to special-case the simulator case all over the framework, just use
getuid and getpid, and intercept those in the simulator wrapper.
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| c5b2c0bf8007562536b822eb060fc54a01f8e08b |
|
20-May-2009 |
Mathias Agopian <mathias@google.com> |
move libbinder's header files under includes/binder
/frameworks/native/libs/binder/IPCThreadState.cpp
|
| 208059f67ed2dd9fa025e07fcb6954d3cb61c79e |
|
19-May-2009 |
Mathias Agopian <mathias@google.com> |
checkpoint: split libutils into libutils + libbinder
/frameworks/native/libs/binder/IPCThreadState.cpp
|