OSDN Git Service
Elliott Hughes [Wed, 30 Jan 2013 00:41:23 +0000 (00:41 +0000)]
Merge "Fix x86 build to use <elf.h>."
Elliott Hughes [Wed, 30 Jan 2013 00:40:39 +0000 (16:40 -0800)]
Fix x86 build to use <elf.h>.
Change-Id: I7b697d5eae69dc08eb31471a42cb8bbe5360be76
Elliott Hughes [Wed, 30 Jan 2013 00:33:53 +0000 (00:33 +0000)]
Merge "Bring the NOTICE files back up to date."
Elliott Hughes [Wed, 30 Jan 2013 00:25:06 +0000 (16:25 -0800)]
Bring the NOTICE files back up to date.
Change-Id: I978275cf6604b90595ee79d897c0460eeadc7dc8
Elliott Hughes [Wed, 30 Jan 2013 00:06:11 +0000 (00:06 +0000)]
Merge "Use the NetBSD <sys/exec_elf.h>."
Elliott Hughes [Tue, 29 Jan 2013 23:02:50 +0000 (15:02 -0800)]
Use the NetBSD <sys/exec_elf.h>.
Replace a kernel header file dependency with files from NetBSD.
They're more complete, and ELF is ELF, whether you're on Linux or a BSD.
Bug:
7973611
Change-Id: I83ee719e7efdf432ec2ddbe8be271d05b2f558d7
Elliott Hughes [Tue, 29 Jan 2013 19:59:30 +0000 (19:59 +0000)]
Merge "Don't collect useless stack frames; do demangle C++ symbols."
Elliott Hughes [Tue, 29 Jan 2013 00:27:36 +0000 (16:27 -0800)]
Don't collect useless stack frames; do demangle C++ symbols.
Previously, we'd collect every stack frame and then throw some away
when we came to log them. This meant that stack traces were effectively
shorter than the buffers that had been allocated for them. This patch
only stores frames we'll actually output.
Also dynamically call the C++ demangler so we don't have to try to
read mangled names. Because no one knows the mangling of operator new[]
for int arrays off the top of their head.
Bug:
7291287
Change-Id: I42b022fd7cd61675d05171de4c3b2704d058ef2a
Elliott Hughes [Mon, 28 Jan 2013 22:04:44 +0000 (22:04 +0000)]
Merge "Unit tests for formatting code, fix %%."
Elliott Hughes [Mon, 28 Jan 2013 18:36:31 +0000 (10:36 -0800)]
Unit tests for formatting code, fix %%.
Also fix <signal.h> and <stdio.h> so they don't cause compiler warnings.
Change-Id: Ib1a746bf01de22d47dbd964de0e6af80a7c96303
Elliott Hughes [Sat, 26 Jan 2013 01:55:43 +0000 (01:55 +0000)]
Merge "More debug malloc fixes."
Elliott Hughes [Sat, 26 Jan 2013 01:13:45 +0000 (17:13 -0800)]
More debug malloc fixes.
Include the leaky executable's name in the log output. Fix the "sh" test.
Use uintptr_t instead of intptr_t.
Also fix debug formatting of NULL with %s.
Bug:
7291287
Change-Id: I015bf341cd48d43a247173612e6ccb1bf1243d53
Elliott Hughes [Fri, 25 Jan 2013 21:57:58 +0000 (21:57 +0000)]
Merge "arm syscall : for eabi call_default don't use stack"
Nick Kralevich [Fri, 25 Jan 2013 21:52:00 +0000 (21:52 +0000)]
Merge "system_properties: do more checking of file"
Nick Kralevich [Fri, 25 Jan 2013 21:07:31 +0000 (13:07 -0800)]
system_properties: do more checking of file
Check that the permissions on the properties file
are exactly as we expect them to be.
Make sure we close the fd if fstat fails.
Refactor the code slightly.
Change-Id: I5503fd58c3b8093ce7e6d05920748ed70eaf8e2c
Nick Kralevich [Thu, 24 Jan 2013 16:44:59 +0000 (16:44 +0000)]
Merge "prctl.h: include sys/cdefs.h"
Nick Kralevich [Thu, 24 Jan 2013 00:49:47 +0000 (16:49 -0800)]
prctl.h: include sys/cdefs.h
prctl.h uses __BEGIN_DECLS but fails to include sys/cdefs.h
(where it's defined). Code which includes prctl.h without
previously including sys/cdefs.h will fail to compile.
Fixed.
Change-Id: If4c9f3308f08b93596dcd00e351ae786807e9320
Nick Kralevich [Wed, 23 Jan 2013 18:50:24 +0000 (18:50 +0000)]
Merge "libc: use more secure system properties if available"
Nick Kralevich [Wed, 23 Jan 2013 17:28:35 +0000 (09:28 -0800)]
libc: use more secure system properties if available
Currently, system properties are passed via the environment
variable ANDROID_PROPERTY_WORKSPACE and a file descriptor passed
from parent to child. This is insecure for setuid executables,
as the environment variable can be changed by the caller.
Modify system property handling so that we get the properties
from a root owned properties file, rather than using an
environment variable. Fall back to the environment variable
if the file doesn't exist.
Bug:
8045561
Change-Id: I54f3efa98cf7d63d88788da5ce0d19e34fd7851a
Elliott Hughes [Wed, 23 Jan 2013 06:06:21 +0000 (06:06 +0000)]
Merge "Don't free anything when reporting leaks."
Elliott Hughes [Wed, 23 Jan 2013 06:05:18 +0000 (06:05 +0000)]
Merge "Disable leak checking for mksh; it's way too leaky."
Elliott Hughes [Wed, 23 Jan 2013 02:36:28 +0000 (18:36 -0800)]
Don't free anything when reporting leaks.
We don't know that they're not going to be cleaned up by a
C++ global destructor that runs after us. This is the case with
bootanimation, for example.
Bug:
7291287
Change-Id: Iba402514d1735fdc2ae4bc95b65396d816be46c0
Elliott Hughes [Wed, 23 Jan 2013 02:35:14 +0000 (18:35 -0800)]
Disable leak checking for mksh; it's way too leaky.
When each shell leaks ~240 allocations, you can't see the leaks from
the program you ran with "adb shell".
Bug:
7291287
Change-Id: Ib8780db72ba0114ebdb24768537da74bbb61f354
Elliott Hughes [Wed, 23 Jan 2013 01:11:28 +0000 (01:11 +0000)]
Merge "Add const for first argument of sigismember for fit POSIX spec"
Elliott Hughes [Tue, 22 Jan 2013 23:19:35 +0000 (23:19 +0000)]
Merge "Our strcoll(3) is no different from NetBSD's, so take exactly theirs."
Elliott Hughes [Tue, 22 Jan 2013 23:10:19 +0000 (15:10 -0800)]
Our strcoll(3) is no different from NetBSD's, so take exactly theirs.
Change-Id: I45251047202a229f9175735ecc23c0ebcda71e8d
Elliott Hughes [Tue, 22 Jan 2013 22:53:53 +0000 (22:53 +0000)]
Merge "Avoid overflow in memccpy."
Elliott Hughes [Tue, 22 Jan 2013 22:41:23 +0000 (14:41 -0800)]
Avoid overflow in memccpy.
Just take the upstream NetBSD code.
Bug: http://code.google.com/p/android/issues/detail?id=43078
Change-Id: Ibbbde9d00e8bc6a09c9503aab2b04b4e3d1f98b0
Elliott Hughes [Tue, 22 Jan 2013 22:37:27 +0000 (22:37 +0000)]
Merge "Clean up debuggerd-related logging."
Elliott Hughes [Tue, 22 Jan 2013 22:26:35 +0000 (22:26 +0000)]
Merge "Use the new non-allocating logging for dlmalloc failures."
Elliott Hughes [Tue, 22 Jan 2013 22:00:09 +0000 (14:00 -0800)]
Use the new non-allocating logging for dlmalloc failures.
Change-Id: I88afe0201ee5766a295fc5a9e710fba9d6e0d363
Elliott Hughes [Tue, 22 Jan 2013 22:17:14 +0000 (14:17 -0800)]
Clean up debuggerd-related logging.
Bug:
7291287
Change-Id: Ia7aa386e8b75b8058d7d9e707e11b1da7dc62f00
Nick Kralevich [Tue, 22 Jan 2013 21:30:27 +0000 (21:30 +0000)]
Merge "bionic_auxv.h: fix #define conflict"
Nick Kralevich [Tue, 22 Jan 2013 21:09:04 +0000 (13:09 -0800)]
bionic_auxv.h: fix #define conflict
Both libc/include/sys/auxv.h and libc/private/bionic_auxv.h
use _SYS_AUXV_H_ to see if a header file has already been included.
This prevents both of these files from being included at the same
time.
Fix this name conflict.
Change-Id: Ifaec88aa9779d784b95f8e75145117acf3d5cfc5
Elliott Hughes [Tue, 22 Jan 2013 19:22:46 +0000 (19:22 +0000)]
Merge "Add missing extern "C"."
Elliott Hughes [Tue, 22 Jan 2013 19:20:45 +0000 (11:20 -0800)]
Add missing extern "C".
Change-Id: Idbf24ce8482ff03f24caa89bafb08677b1c5cec3
Elliott Hughes [Tue, 22 Jan 2013 17:44:15 +0000 (17:44 +0000)]
Merge "Fix the duplication in the debugging code."
Kito Cheng [Sat, 19 Jan 2013 16:15:15 +0000 (00:15 +0800)]
Add const for first argument of sigismember for fit POSIX spec
Change-Id: Icbc67375282f2a22dce02e4bacab15ddae846057
Elliott Hughes [Fri, 18 Jan 2013 02:36:06 +0000 (18:36 -0800)]
Fix the duplication in the debugging code.
We had two copies of the backtrace code, and two copies of the
libcorkscrew /proc/pid/maps code. This patch gets us down to one.
We also had hacks so we could log in the malloc debugging code.
This patch pulls the non-allocating "printf" code out of the
dynamic linker so everyone can share.
This patch also makes the leak diagnostics easier to read, and
makes it possible to paste them directly into the 'stack' tool (by
using relative PCs).
This patch also fixes the stdio standard stream leak that was
causing a leak warning every time tf_daemon ran.
Bug:
7291287
Change-Id: I66e4083ac2c5606c8d2737cb45c8ac8a32c7cfe8
Nick Kralevich [Fri, 18 Jan 2013 20:50:40 +0000 (20:50 +0000)]
Merge "Revert "Filter ANDROID_PROPERTY_WORKSPACE""
Guang Zhu [Fri, 18 Jan 2013 07:25:24 +0000 (23:25 -0800)]
Revert "Filter ANDROID_PROPERTY_WORKSPACE"
Temporarily revert the change since DNS resolution seems
broken right now in ping util.
Bug:
8029617
This reverts commit
a0f64756a4a55ab48b2b5511d4e7c45583dac44b.
Nick Kralevich [Fri, 18 Jan 2013 00:49:35 +0000 (00:49 +0000)]
Merge "FORTIFY_SOURCE: optimize"
Elliott Hughes [Fri, 18 Jan 2013 00:34:13 +0000 (00:34 +0000)]
Merge "Add functionlity to the scripts to replace tokens in kernel headers based on architecture."
Nick Kralevich [Thu, 17 Jan 2013 23:41:33 +0000 (15:41 -0800)]
FORTIFY_SOURCE: optimize
Don't do the fortify_source checks if we can determine, at
compile time, that the provided operation is safe.
This avoids silliness like calling fortify source on things like:
size_t len = strlen("asdf");
printf("%d\n", len);
and allows the compiler to optimize this code to:
printf("%d\n", 4);
Defer to gcc's builtin functions instead of pointing our code
to the libc implementation.
Change-Id: I5e1dcb61946461c4afaaaa983e39f07c7a0df0ae
Raghu Gandham [Thu, 17 Jan 2013 00:42:47 +0000 (16:42 -0800)]
Add functionlity to the scripts to replace tokens in kernel headers
based on architecture.
Nick Kralevich [Thu, 17 Jan 2013 16:41:31 +0000 (16:41 +0000)]
Merge "linker: add -Wl,--exclude-libs,ALL to LDFLAGS"
Nick Kralevich [Thu, 17 Jan 2013 00:43:15 +0000 (16:43 -0800)]
linker: add -Wl,--exclude-libs,ALL to LDFLAGS
The linker is essentially a shared library, and incorporates
it's own copy of libc. Even though it's unnecessary, currently
/system/bin/linker is exporting various libc symbols (only to
apps which explicitly dlopen /system/bin/linker)
Add --exclude-libs,ALL, which tells the static linker to mark
all of the imported libc symbols as hidden. This reduces the
size of /system/bin/linker from 92K to 67K with no obvious
loss in functionality.
$ adb shell ls -l /system/bin/linker
-rwxrwxrwx root root 92260 2013-01-16 16:52 linker
$ adb shell ls -l /system/bin/linker
-rwxrwxrwx root root 67660 2013-01-16 16:49 linker
Documentation on exclude-libs can be found at
http://sourceware.org/binutils/docs-2.21/ld/Options.html
Change-Id: I4508287770e4b7a845def2e6b4af969f9c866c6a
Nick Kralevich [Wed, 16 Jan 2013 21:36:53 +0000 (21:36 +0000)]
Merge "Revert "stack protector: use AT_RANDOM""
Nick Kralevich [Wed, 16 Jan 2013 21:13:22 +0000 (13:13 -0800)]
Revert "stack protector: use AT_RANDOM"
The AT_RANDOM changes broke setuid / setgid executables
such as "ping". When the linker executes a setuid program,
it cleans the environment, removing any invalid environment
entries, and adding "NULL"s to the end of the environment
array for each removed variable. Later on, we try to determine
the location of the aux environment variable, and get tripped
up by these extra NULLs.
Reverting this patch will get setuid executables working again,
but getauxval() is still broken for setuid programs because of
this bug.
This reverts commit
e3a49a8661125f24aec8a1453e54b3b78005e21e.
Change-Id: I05c58a896b1fe32cfb5d95d43b096045cda0aa4a
Nick Kralevich [Wed, 16 Jan 2013 19:23:25 +0000 (11:23 -0800)]
Merge "stack protector: use AT_RANDOM"
Nick Kralevich [Wed, 16 Jan 2013 19:11:17 +0000 (11:11 -0800)]
Merge "Filter ANDROID_PROPERTY_WORKSPACE"
Nick Kralevich [Mon, 14 Jan 2013 22:46:26 +0000 (14:46 -0800)]
stack protector: use AT_RANDOM
Populate the stack canaries from the kernel supplied
AT_RANDOM value, which doesn't involve any system calls.
This is slightly faster (6 fewer syscalls) and avoids
unnecessarily reading /dev/urandom, which depletes entropy.
Bug:
7959813
Change-Id: If2b43100a2a9929666df3de56b6139fed969e0f1
Elliott Hughes [Wed, 16 Jan 2013 17:42:51 +0000 (09:42 -0800)]
Merge "Fix signalfd for MIPS."
Elliott Hughes [Wed, 16 Jan 2013 17:34:16 +0000 (09:34 -0800)]
Fix signalfd for MIPS.
Also mark signalfd's sigset_t* argument as non-nullable.
Change-Id: I466e09cdf3fb92480744c496da92274a97f99dd1
Matthieu Castet [Wed, 16 Jan 2013 13:02:50 +0000 (14:02 +0100)]
arm syscall : for eabi call_default don't use stack
In the default case, we don't need to use the stack, we can save r7 with
ip register (that what does eglibc).
This allow to fix vfork data corruption
(see
3884bfe9661955543ce203c60f9225bbdf33f6bb), because vfork now don't
use the stack.
Nick Kralevich [Wed, 16 Jan 2013 00:02:03 +0000 (16:02 -0800)]
Filter ANDROID_PROPERTY_WORKSPACE
When executing a setuid executable, filter out ANDROID_PROPERTY_WORKSPACE
from the environment. Some applications implicitly trust the property
space and don't realize that it's passed by an environment variable
which can be modified by the caller.
Change-Id: I3e3a98941f0a1f249a2ff983ecbcfe1278aa9159
Elliott Hughes [Tue, 15 Jan 2013 21:21:51 +0000 (13:21 -0800)]
Merge "Revert "DO NOT MERGE Revert "Add the libcutils localtime_tz and mktime_t extensions to bionic."""
Elliott Hughes [Tue, 15 Jan 2013 19:12:18 +0000 (11:12 -0800)]
Revert "DO NOT MERGE Revert "Add the libcutils localtime_tz and mktime_t extensions to bionic.""
This reverts commit
f4b34b6c3942be273ad7298a40be0d312b183aab.
The revert was only meant to apply to the jb-mr1 branch, but accidentally
leaked out into AOSP. This revert-revert gets AOSP master and internal
master back in sync.
Nick Kralevich [Tue, 15 Jan 2013 18:56:06 +0000 (10:56 -0800)]
Merge "fix strerror_r test"
Nick Kralevich [Tue, 15 Jan 2013 18:35:09 +0000 (10:35 -0800)]
fix strerror_r test
e6e60065ff093ff8c859ab146cf543531cb1967c modified strerror_r to
treat errno as signed. However, the change to the test code
modified the "strerror" test, not the "strerror_r" test.
Make the same change for the strerror_r code.
Change-Id: Ia236a53df5745935e229a4446a74da8bed0cfd7b
Ben Cheng [Mon, 14 Jan 2013 23:35:03 +0000 (15:35 -0800)]
Merge "Add __aeabi_idiv to the dummy reference list."
Ben Cheng [Mon, 14 Jan 2013 19:26:51 +0000 (11:26 -0800)]
Add __aeabi_idiv to the dummy reference list.
If the platform code is compiled with -mcpu=cortex-a15, then without this
change prebuilt libraries built against -march=armv7 cannot resolve the
dependency on __aeabi_idiv (provided by libgcc.a).
Bug:
7961327
cherry-picked from internal master.
Change-Id: I8fe59a98eb53d641518b882523c1d6a724fb7e55
Nick Kralevich [Mon, 14 Jan 2013 22:14:49 +0000 (14:14 -0800)]
Merge "headers: update auxvec.h from Linux kernel"
Nick Kralevich [Mon, 14 Jan 2013 19:28:26 +0000 (11:28 -0800)]
headers: update auxvec.h from Linux kernel
Pull a new version of auxvec.h from the upstream Linux
kernel at commit
b719f43059903820c31edb30f4663a2818836e7f
These files were generated using the following commands:
cd bionic/libc/kernel
./tools/clean_header.py -u ../../../external/kernel-headers/original/uapi/linux/auxvec.h
./tools/clean_header.py -u ../../../external/kernel-headers/original/linux/auxvec.h
./tools/clean_header.py -u ../../../external/kernel-headers/original/asm-x86/auxvec.h
This change is needed to get AT_RANDOM defined.
Change-Id: Ib064649684b17af6ff4b1a31d501a05f78bb81d0
Ian Rogers [Mon, 14 Jan 2013 18:54:44 +0000 (10:54 -0800)]
Merge "Name anonymous mmap mallocs."
Nick Kralevich [Mon, 14 Jan 2013 18:21:23 +0000 (10:21 -0800)]
Merge "libc_init_static: apply relro earlier."
Elliott Hughes [Mon, 14 Jan 2013 17:57:45 +0000 (09:57 -0800)]
Merge "Fix my git mistake."
Elliott Hughes [Mon, 14 Jan 2013 17:56:21 +0000 (09:56 -0800)]
Fix my git mistake.
This was the formatting change that was supposed to be in
cf23905a4bcc7bfdd109be5b6d69ad06877aa217.
Change-Id: Ib79fa031b68f6f541f532507eb589afeaedb831f
Elliott Hughes [Mon, 14 Jan 2013 17:31:04 +0000 (09:31 -0800)]
Merge "[MIPS] Set DT_DEBUG dyntab entry if it is writable"
Chris Dearman [Fri, 11 Jan 2013 23:32:20 +0000 (15:32 -0800)]
[MIPS] Set DT_DEBUG dyntab entry if it is writable
This is primarily for MIPS exutables that do not have a
DT_MIPS_RLD_MAP entry.
Change-Id: I4c221d92debcfed961eeee2515123f3fb21ec8e6
Signed-off-by: Chris Dearman <chris@mips.com>
Nick Kralevich [Sat, 12 Jan 2013 02:38:26 +0000 (18:38 -0800)]
libc_init_static: apply relro earlier.
The dynamic linker applies relro before the preinit and init
arrays are executed, so we should be consistent for statically
linked executables.
Change-Id: Ia0a49d0e981a6e8791f74eed00280edf576ba139
Ian Rogers [Fri, 11 Jan 2013 23:25:44 +0000 (15:25 -0800)]
Name anonymous mmap mallocs.
Change-Id: Icc53ba1eecb8445210623826d8e99a611d686f7f
Nick Kralevich [Fri, 11 Jan 2013 22:43:05 +0000 (14:43 -0800)]
libc: add getauxval()
Add support for getauxval(). This method allows a program an easy way
to retrieve information from the kernel auxiliary vector, and will
hopefully replace other clumsy ways of accessing this same information.
This particular function was also added to glibc in glibc 2.16.
See the following URLs for more details.
* http://lwn.net/Articles/519085/
* http://www.gnu.org/software/libc/manual/html_node/Auxiliary-Vector.html
This change is a prerequisite for bug
7959813.
Bug: http://code.google.com/p/android/issues/detail?id=38441
Change-Id: Iba19d899df334bddc6f4899077ece2fc87564ea8
Nick Kralevich [Fri, 11 Jan 2013 19:03:40 +0000 (11:03 -0800)]
Merge "Add stack canaries / strcpy tests."
Nick Kralevich [Fri, 11 Jan 2013 01:12:29 +0000 (17:12 -0800)]
Add stack canaries / strcpy tests.
Add a test to ensure that stack canaries are working
correctly. Since stack canaries aren't normally generated
on non-string functions, we have to enable stack-protector-all.
Add a test to ensure that an out of bounds strcpy generates
a runtime failure.
Change-Id: Id0d3e59fc4b9602da019e4d35c5c653e1a57fae4
Wink Saville [Tue, 8 Jan 2013 23:15:45 +0000 (15:15 -0800)]
Fix unused warnings in pthread.c
Change-Id: I0287aadb825fd8cda29dc976bce55d75a1279fc5
Elliott Hughes [Fri, 11 Jan 2013 00:24:36 +0000 (16:24 -0800)]
Merge "glibc 2.15 treats errno as signed in strerror(3)."
Elliott Hughes [Fri, 11 Jan 2013 00:01:59 +0000 (16:01 -0800)]
glibc 2.15 treats errno as signed in strerror(3).
And the only reason I hadn't done that in bionic is because I wanted to behave
the same as glibc.
Change-Id: I2cf1bf0aac82a748cd6305a2cabbac0790058570
Elliott Hughes [Thu, 10 Jan 2013 23:12:46 +0000 (15:12 -0800)]
Merge "Only have one copy of the kernel_sigset_t hack, and add more tests."
Elliott Hughes [Thu, 10 Jan 2013 22:42:14 +0000 (14:42 -0800)]
Only have one copy of the kernel_sigset_t hack, and add more tests.
Change-Id: I377522fcba6fb4b5fd2754ab15b091014bd7c16f
Elliott Hughes [Thu, 10 Jan 2013 21:17:26 +0000 (13:17 -0800)]
Merge "Add signalfd call to bionic"
Rom Lemarchand [Wed, 9 Jan 2013 23:46:06 +0000 (15:46 -0800)]
Add signalfd call to bionic
Add signalfd() call to bionic.
Adding the signalfd call was done in 3 steps:
- add signalfd4 system call (function name and syscall
number) to libc/SYSCALLS.TXT
- generate all necessary headers by calling
libc/tools/gensyscalls.py. This patch is adding
the generated files since the build system
does not call gensyscalls.py.
- create the signalfd wrapper in signalfd.cpp and add
the function prototype to sys/signalfd.h
(cherry-pick of
0c11611c11f4dc1b6d43587b72c3ccbe8c51a51c, modified to
work with older versions of GCC still in use on some branches.)
Change-Id: I4c6c3f12199559af8be63f93a5336851b7e63355
Elliott Hughes [Mon, 7 Jan 2013 22:46:13 +0000 (14:46 -0800)]
Merge "Don't test GNU-style ELF hashes on MIPS."
Elliott Hughes [Mon, 7 Jan 2013 22:18:22 +0000 (14:18 -0800)]
Don't test GNU-style ELF hashes on MIPS.
The MIPS toolchain can't generate them because they're incompatible
with the MIPS ABI (which requires .dynsym match the GOT, while GNU-style
requires .dynsym to be sorted by hash code), so there's nothing to test.
Change-Id: I2220f452fe6fe595ec1312544cc741dd390a36a5
Elliott Hughes [Mon, 7 Jan 2013 22:11:43 +0000 (14:11 -0800)]
Merge "Fix an off-by-one error in the sigset_t function error handling."
Elliott Hughes [Mon, 7 Jan 2013 21:58:49 +0000 (13:58 -0800)]
Fix an off-by-one error in the sigset_t function error handling.
Spotted while running the tests on MIPS, where sigset_t is
actually large enough. The bits in sigset_t are used such that
signal 1 is represented by bit 0, so the range of signals is
actually [1, 8*sizeof(sigset_t)]; it seems clearer to reword
the code in terms of valid bit offsets [0, 8*sizeof(sigset_t)),
which leads to the usual bounds checking idiom.
Change-Id: Id899c288e15ff71c85dd2fd33c47f8e97aa1956f
Elliott Hughes [Mon, 7 Jan 2013 21:30:58 +0000 (13:30 -0800)]
Merge "[MIPS] Rewrite fenv.h for Android"
Raghu Gandham [Sat, 5 Jan 2013 00:29:14 +0000 (16:29 -0800)]
[MIPS] Rewrite fenv.h for Android
Change-Id: I4d1e2f0b37b587426ccc9f26c525ec0d36637c7d
Elliott Hughes [Fri, 4 Jan 2013 00:55:30 +0000 (16:55 -0800)]
Merge "Add AF_CAN and PF_CAN (and other missing families)."
Elliott Hughes [Fri, 4 Jan 2013 00:25:47 +0000 (16:25 -0800)]
Add AF_CAN and PF_CAN (and other missing families).
Change-Id: I2c183a6f5f7a7e81e87dad85d8c9aff9c43ed33a
Elliott Hughes [Fri, 4 Jan 2013 00:20:01 +0000 (16:20 -0800)]
Merge "Fix debug malloc."
Elliott Hughes [Thu, 3 Jan 2013 23:44:03 +0000 (15:44 -0800)]
Fix debug malloc.
...which has been broken since the linker data structures went read-only.
Bug:
7941716
Change-Id: If28f6bac0fcb13e371e4d85b064544f561c8d692
Elliott Hughes [Thu, 3 Jan 2013 19:54:56 +0000 (11:54 -0800)]
Merge "sysconf.c was renamed to sysconf.cpp (and modified)..."
Elliott Hughes [Thu, 3 Jan 2013 19:53:35 +0000 (11:53 -0800)]
sysconf.c was renamed to sysconf.cpp (and modified)...
...but sysconf.c still lingers on due to some git/repo accident. Kill it.
Change-Id: Iae354ecb21abf03a3f718cc45cfdddb7a9347778
Elliott Hughes [Wed, 2 Jan 2013 23:51:15 +0000 (15:51 -0800)]
Merge "Define _POSIX_MONOTONIC_CLOCK and implement sysconf(_SC_MONOTONIC_CLOCK)."
Elliott Hughes [Wed, 2 Jan 2013 22:23:43 +0000 (14:23 -0800)]
Define _POSIX_MONOTONIC_CLOCK and implement sysconf(_SC_MONOTONIC_CLOCK).
Bug: http://code.google.com/p/android/issues/detail?id=39680
Change-Id: I11cf10a66f9d305868a725f04f581099fb88bbfc
Elliott Hughes [Sat, 22 Dec 2012 01:46:22 +0000 (17:46 -0800)]
Merge "Support System.loadLibrary for libraries with transitive dependencies."
Elliott Hughes [Fri, 21 Dec 2012 03:08:17 +0000 (19:08 -0800)]
Merge "Fix format_number."
Elliott Hughes [Fri, 21 Dec 2012 02:59:05 +0000 (18:59 -0800)]
Fix format_number.
I broke this the other day when silencing x86 gcc warnings.
Bug:
7904160
Change-Id: I8e60cc1f8cbaff95248c8738d84e515413d839e4
Elliott Hughes [Thu, 20 Dec 2012 22:42:14 +0000 (14:42 -0800)]
Support System.loadLibrary for libraries with transitive dependencies.
Also fix the FLAG_ERROR annoyance --- it's not helpful to cache failures.
Bug:
7896159
Bug: http://code.google.com/p/android/issues/detail?id=34416
Bug: http://code.google.com/p/android/issues/detail?id=22143
Change-Id: I60f235edb4ea4756e1f7ce56f7739f18e8a50789