bionic/libc
Jean-Baptiste Queru faca92f2f1 Handle pthread-related changes (mutex/atfork)
First commit:

Revert "Revert "am be741d47: am 2f460fbe: am 73b5cad9: Merge "bionic: Fix wrong kernel_id in pthread descriptor after fork()"""

This reverts commit 06823da2f0.

Second commit:

bionic: fix atfork hanlder_mutex deadlock

This cherry-picks commit 34e89c232d

After applying the kernel_id fix, the system refused to boot up and we
got following crash log:
I/DEBUG   (  113): pid: 618, tid: 618  >>> org.simalliance.openmobileapi.service:remote <<<
I/DEBUG   (  113): signal 16 (SIGSTKFLT), code -6 (?), fault addr --------
I/DEBUG   (  113):  eax fffffe00  ebx b77de994  ecx 00000080  edx 00724002
I/DEBUG   (  113):  esi 00000000  edi 00004000
I/DEBUG   (  113):  xcs 00000073  xds 0000007b  xes 0000007b  xfs 00000000 xss 0000007b
I/DEBUG   (  113):  eip b7761351  ebp bfdf3de8  esp bfdf3dc4  flags 00000202
I/DEBUG   (  113):     #00  eip: 00015351  /system/lib/libc.so
I/DEBUG   (  113):     #01  eip: 0000d13c  /system/lib/libc.so (pthread_mutex_lock)
I/DEBUG   (  113):     #02  eip: 00077b48  /system/lib/libc.so (__bionic_atfork_run_prepare)
I/DEBUG   (  113):     #03  eip: 00052cdb  /system/lib/libc.so (fork)
I/DEBUG   (  113):     #04  eip: 0009ae91  /system/lib/libdvm.so (_Z18dvmOptimizeDexFileillPKcjjb)
I/DEBUG   (  113):     #05  eip: 000819d6  /system/lib/libdvm.so (_Z14dvmJarFileOpenPKcS0_PP7JarFileb)
I/DEBUG   (  113):     #06  eip: 000b175e  /system/lib/libdvm.so (_ZL40Dalvik_dalvik_system_DexFile_openDexFilePKjP6JValue)
I/DEBUG   (  113):     #07  eip: 0011fb94  /system/lib/libdvm.so

Root cause:
The atfork uses the mutex handler_mutex to protect the atfork_head. The
parent will call __bionic_atfork_run_prepare() to lock the handler_mutex,
and need both the parent and child to unlock their own copy of handler_mutex
after fork. At that time, the owner of hanlder_mutex is set as the parent.
If we apply the kernel_id fix, then the child's kernel_id will be set as
child's tid.
The handler_mutex is a recursive lock, and pthread_mutex_unlock(&hander_mutex)
will fail because the mutex owner is the parent, while the current tid
(__get_thread()->kernel_id) is child, not matched with the mutex owner.
At that time, the handler_mutex is left in lock state.If the child wants to
fork other process after than, then it will try to lock handler_mutex, and
then be deadlocked.

Fix:
Since the child has its own copy of vm space from the the parent, the
child space's handler_mutex should be reset to the initialized state.

Change-Id: I3907dd9a153418fb78862f2aa6d0302c375d9e27
Signed-off-by: Jack Ren <jack.ren@intel.com>
Signed-off-by: Chenyang Du <chenyang.du@intel.com>
Signed-off-by: Bruce Beare <bruce.j.beare@intel.com>

Change-Id: Ic8072f366a877443a60fe215f3c00b3df5a259c8
2012-03-26 15:38:59 -07:00
..
arch-arm Update kernel headers and add syscall "perf_event_open" 2012-03-13 12:28:40 -07:00
arch-sh/syscalls Merge c4cb87f3 2012-02-01 09:46:08 -08:00
arch-x86 Update kernel headers and add syscall "perf_event_open" 2012-03-13 12:28:40 -07:00
bionic Handle pthread-related changes (mutex/atfork) 2012-03-26 15:38:59 -07:00
docs libc: Fix the definition of SIGRTMAX 2010-12-20 15:58:06 +01:00
include Merge "string.h: add __attribute__ ((pure)) to string functions" 2012-03-21 14:40:41 -07:00
inet Fix build. 2011-06-09 13:03:17 -07:00
kernel Fixed to #include correct 32-bit headers; Refreshed libc/kernel headers 2012-03-23 23:07:36 +08:00
netbsd Use new binary code format 2012-03-09 11:50:46 -08:00
private Handle pthread-related changes (mutex/atfork) 2012-03-26 15:38:59 -07:00
regex Remove compiler warnings when building Bionic. 2010-06-22 17:51:41 -07:00
stdio libc: speed-up flockfile()/funlockfile() 2011-11-15 13:16:42 +01:00
stdlib Enable functional DSO object destruction 2011-07-07 22:51:43 +02:00
string string: Fix wrong comparison semantics 2011-12-05 18:37:10 -08:00
tools Merge c4cb87f3 2012-02-01 09:46:08 -08:00
tzcode libc: remove private declarations from <time.h> and <resolv.h> 2012-01-13 14:24:08 +01:00
unistd Merge "bionic: pass MADV_MERGEABLE on private & anonymous mmaps" 2012-03-21 13:51:42 -07:00
wchar wchar.h: improve wchar_t support in Bionic 2010-06-15 07:04:41 -07:00
zoneinfo Upgrade to tzdata2012b. 2012-03-01 23:34:11 -08:00
Android.mk New additions/bug fixes required/found when porting perf. 2012-03-20 12:54:55 -07:00
CAVEATS auto import from //depot/cupcake/@135843 2009-03-03 19:28:35 -08:00
Jamfile auto import from //depot/cupcake/@135843 2009-03-03 19:28:35 -08:00
MODULE_LICENSE_BSD auto import from //depot/cupcake/@135843 2009-03-03 19:28:35 -08:00
NOTICE Clean up NOTICE files. 2010-10-19 15:12:40 -07:00
README Add an 's and a . to the bionic/libc README. 2009-07-23 17:41:47 -07:00
SYSCALLS.TXT Update kernel headers and add syscall "perf_event_open" 2012-03-13 12:28:40 -07:00

Welcome to Bionic, Android's small and custom C library for the Android
platform.

Bionic is mainly a port of the BSD C library to our Linux kernel with the
following additions/changes:

- no support for locales
- no support for wide chars (i.e. multi-byte characters)
- its own smallish implementation of pthreads based on Linux futexes
- support for x86, ARM and ARM thumb CPU instruction sets and kernel interfaces

Bionic is released under the standard 3-clause BSD License

Bionic doesn't want to implement all features of a traditional C library, we only
add features to it as we need them, and we try to keep things as simple and small
as possible. Our goal is not to support scaling to thousands of concurrent threads
on multi-processors machines; we're running this on cell-phones, damnit !!

Note that Bionic doesn't provide a libthread_db or a libm implementation.


Adding new syscalls:
====================

Bionic provides the gensyscalls.py Python script to automatically generate syscall
stubs from the list defined in the file SYSCALLS.TXT. You can thus add a new syscall
by doing the following:

- edit SYSCALLS.TXT
- add a new line describing your syscall, it should look like:

   return_type  syscall_name(parameters)    syscall_number

- in the event where you want to differentiate the syscall function from its entry name,
  use the alternate:

   return_type  funcname:syscall_name(parameters)  syscall_number

- additionally, if the syscall number is different between ARM and x86, use:

   return_type  funcname[:syscall_name](parameters)   arm_number,x86_number

- a syscall number can be -1 to indicate that the syscall is not implemented on
  a given platform, for example:

   void   __set_tls(void*)   arm_number,-1


the comments in SYSCALLS.TXT contain more information about the line format

You can also use the 'checksyscalls.py' script to check that all the syscall
numbers you entered are correct. It does so by looking at the values defined in
your Linux kernel headers. The script indicates where the values are incorrect
and what is expected instead.