OSDN Git Service

android-x86/bionic.git
9 years agoAllow wildcards to match arg values.
Christopher Ferris [Sat, 21 Feb 2015 02:31:06 +0000 (18:31 -0800)]
Allow wildcards to match arg values.

Change-Id: I38230b500bb8f8f69af0d7c740855a401cd12898

9 years agoMerge "Move getusershell/endusershell/setusershell to ndk_cruft.cpp."
Yabin Cui [Wed, 25 Feb 2015 00:00:51 +0000 (00:00 +0000)]
Merge "Move getusershell/endusershell/setusershell to ndk_cruft.cpp."

9 years agoMerge "Make all output of child test go to parent process in gtest_main."
Yabin Cui [Tue, 24 Feb 2015 23:40:30 +0000 (23:40 +0000)]
Merge "Make all output of child test go to parent process in gtest_main."

9 years agoMove getusershell/endusershell/setusershell to ndk_cruft.cpp.
Yabin Cui [Tue, 24 Feb 2015 21:18:25 +0000 (13:18 -0800)]
Move getusershell/endusershell/setusershell to ndk_cruft.cpp.

Bug: 19108648
Change-Id: I65134040345398dde612c4325d902db58f3bd2be

9 years agoMake all output of child test go to parent process in gtest_main.
Yabin Cui [Tue, 24 Feb 2015 22:39:19 +0000 (14:39 -0800)]
Make all output of child test go to parent process in gtest_main.

Change-Id: Iad460e89755051cdb99593cbf42c97d9a359f32b

9 years agoMerge "Fix "faccessat ignores flags""
Nick Kralevich [Tue, 24 Feb 2015 21:49:09 +0000 (21:49 +0000)]
Merge "Fix "faccessat ignores flags""

9 years agoFix "faccessat ignores flags"
Nick Kralevich [Tue, 24 Feb 2015 21:40:43 +0000 (13:40 -0800)]
Fix "faccessat ignores flags"

The kernel system call faccessat() does not have any flags arguments,
so passing flags to the kernel is currently ignored.

Fix the kernel system call so that no flags argument is passed in.

Ensure that we don't support AT_SYMLINK_NOFOLLOW. This non-POSIX
(http://pubs.opengroup.org/onlinepubs/9699919799/functions/access.html)
flag is a glibc extension, and has non-intuitive, error prone behavior.

For example, consider the following code:

  symlink("foo.is.dangling", "foo");
  if (faccessat(AT_FDCWD, "foo", R_OK, AT_SYMLINK_NOFOLLOW) == 0) {
    int fd = openat(AT_FDCWD, "foo", O_RDONLY | O_NOFOLLOW);
  }

The faccessat() call in glibc will return true, but an attempt to
open the dangling symlink will end up failing. GLIBC documents this
as returning the access mode of the symlink itself, which will
always return true for any symlink on Linux.

Some further discussions of this are at:

  * http://lists.landley.net/pipermail/toybox-landley.net/2014-September/003617.html
  * http://permalink.gmane.org/gmane.linux.lib.musl.general/6952

AT_SYMLINK_NOFOLLOW seems broken by design. I suspect this is why this
function was never added to POSIX. (note that "access" is pretty much
broken by design too, since it introduces a race condition between
check and action). We shouldn't support this until it's clearly
documented by POSIX or we can have it produce intuitive results.

Don't support AT_EACCESS for now. Implementing it is complicated, and
pretty much useless on Android, since we don't have setuid binaries.
See http://git.musl-libc.org/cgit/musl/commit/?id=0a05eace163cee9b08571d2ff9d90f5e82d9c228
for how an implementation might look.

Bug: 18867827
Change-Id: I25b86c5020f3152ffa3ac3047f6c4152908d0e04

9 years agoMerge "Fix dup2 in the case where the two fds are equal."
Elliott Hughes [Tue, 24 Feb 2015 05:55:37 +0000 (05:55 +0000)]
Merge "Fix dup2 in the case where the two fds are equal."

9 years agoFix dup2 in the case where the two fds are equal.
Elliott Hughes [Tue, 24 Feb 2015 02:02:29 +0000 (18:02 -0800)]
Fix dup2 in the case where the two fds are equal.

dup3's behavior differs from dup2 in this case, so we need to paper
over that in the C library.

Change-Id: I313cd6f226db5e237f61866f324c5ecdd12bf762

9 years agoMerge "Add RLIM_SAVED_CUR and RLIM_SAVED_MAX."
Elliott Hughes [Mon, 23 Feb 2015 22:36:06 +0000 (22:36 +0000)]
Merge "Add RLIM_SAVED_CUR and RLIM_SAVED_MAX."

9 years agoAdd RLIM_SAVED_CUR and RLIM_SAVED_MAX.
Elliott Hughes [Mon, 23 Feb 2015 21:41:40 +0000 (13:41 -0800)]
Add RLIM_SAVED_CUR and RLIM_SAVED_MAX.

Change-Id: Ia6be76460bfcf852832325c5f36cb272f49a4b87

9 years agoMerge "[MIPS64] Fix mips64 build."
Elliott Hughes [Mon, 23 Feb 2015 16:04:20 +0000 (16:04 +0000)]
Merge "[MIPS64] Fix mips64 build."

9 years ago[MIPS64] Fix mips64 build.
Nikola Veljkovic [Mon, 23 Feb 2015 15:14:56 +0000 (16:14 +0100)]
[MIPS64] Fix mips64 build.

Build was broken by:
https://android-review.googlesource.com/133834

Use <unistd.h> to get syscall().
Remove <asm/unistd.h>, it gets included through <sys/syscall.h>.

Change-Id: Id762f6dea5f9538c19b79cdd46deda978efd50fe

9 years agoMerge "Fix 64-bit benchmark build"
Greg Hackmann [Fri, 20 Feb 2015 19:16:48 +0000 (19:16 +0000)]
Merge "Fix 64-bit benchmark build"

9 years agoFix 64-bit benchmark build
Greg Hackmann [Fri, 20 Feb 2015 19:00:14 +0000 (11:00 -0800)]
Fix 64-bit benchmark build

The * flag to printf() wants an int instead of size_t, and these are
distinct types on 64-bit.  To accomodate this, make the name column
width helpers return int.

In theory this truncates things, but in practice this only matters if
you have a benchmark with more than INT_MAX characters in its name (in
which case you have bigger problems).

Change-Id: I3338948c25a3a8d84f1ead2f5b457c05da8a01cf
Signed-off-by: Greg Hackmann <ghackmann@google.com>
9 years agoMerge "valgrind can't find syscall because we didn't put it in <unistd.h>."
Elliott Hughes [Fri, 20 Feb 2015 16:56:57 +0000 (16:56 +0000)]
Merge "valgrind can't find syscall because we didn't put it in <unistd.h>."

9 years agovalgrind can't find syscall because we didn't put it in <unistd.h>.
Elliott Hughes [Fri, 20 Feb 2015 06:49:44 +0000 (22:49 -0800)]
valgrind can't find syscall because we didn't put it in <unistd.h>.

Change-Id: I1e47291d4476bd2816138a8cf58f29d4986d39e3

9 years agoMerge "Refactor the benchmark code."
Christopher Ferris [Fri, 20 Feb 2015 00:53:24 +0000 (00:53 +0000)]
Merge "Refactor the benchmark code."

9 years agoRefactor the benchmark code.
Christopher Ferris [Wed, 18 Feb 2015 03:58:53 +0000 (19:58 -0800)]
Refactor the benchmark code.

Changes:
- Modify the benchmarks to derive from a single Benchmark object.
- Rewrite the main iteration code. This includes changing the iteration
  code to use the actual total time calculated by the benchmark as a basis
  for determining whether there are enough iterations instead of using
  the time it takes to run the benchmark.
- Allow benchmarks to take no argument, int, or double.
- Fix the PrettyInt printer for negative integers.
- Modify the max column width name to include the whole name including
  the arg part.
- Reformat property_benchmark.cpp in line with the rest of the code.
- Modify a few of the math benchmarks to take an argument instead of
  separate benchmarks for the same function with different args.
- Create a vector of regex_t structs to represent the args all at
  once instead of when running each benchmark.

This change is in preparation for adding new math based benchmarks.

Tested by running on a nexus flo running at max using the new code
and the old code and comparing. All of the numbers are similar, but
some of the iterations are different due to the slightly different
algorithm used.

Change-Id: I57ad1f3ff083282b9ffeb72e687cab369ce3523a

9 years agoMerge "Add GNU extensions mempcpy and wmemcpy."
Elliott Hughes [Thu, 19 Feb 2015 15:49:53 +0000 (15:49 +0000)]
Merge "Add GNU extensions mempcpy and wmemcpy."

9 years agoAdd GNU extensions mempcpy and wmemcpy.
Elliott Hughes [Thu, 19 Feb 2015 05:29:13 +0000 (21:29 -0800)]
Add GNU extensions mempcpy and wmemcpy.

Used by elfutils. On the bright side, they stopped using __mempcpy.

Bug: 18374026
Change-Id: Id29bbe6ef1c5ed5a171bb6c32182f129d8332abb

9 years agoMerge "Parameterize use of clang in libc_ndk."
Dan Albert [Thu, 19 Feb 2015 02:33:08 +0000 (02:33 +0000)]
Merge "Parameterize use of clang in libc_ndk."

9 years agoParameterize use of clang in libc_ndk.
Dan Albert [Thu, 19 Feb 2015 01:37:52 +0000 (17:37 -0800)]
Parameterize use of clang in libc_ndk.

We still have issues with clang coverage in static libraries, so we
need to make sure we follow suit with the rest of libc for now.

Bug: 17574078
Change-Id: I2ab58a84b1caa0d8d08415d240c35adec5b1e150

9 years agoMerge "[MIPS] Fix setjmp signals"
Elliott Hughes [Thu, 19 Feb 2015 01:00:18 +0000 (01:00 +0000)]
Merge "[MIPS] Fix setjmp signals"

9 years agoMerge "Make gets() deprecated."
Yabin Cui [Wed, 18 Feb 2015 21:33:01 +0000 (21:33 +0000)]
Merge "Make gets() deprecated."

9 years agoMake gets() deprecated.
Yabin Cui [Wed, 18 Feb 2015 20:20:32 +0000 (12:20 -0800)]
Make gets() deprecated.

gets is already deprecated in glibc. Others with __warnattr are not deprecated.

Change-Id: I80a276d2b5964630218be47f1c94b146c0d31151

9 years agoMerge "Implement _FILE_OFFSET_BITS (mostly)."
Elliott Hughes [Wed, 18 Feb 2015 18:01:04 +0000 (18:01 +0000)]
Merge "Implement _FILE_OFFSET_BITS (mostly)."

9 years agoMerge "Update tzdata to tzdata2015a"
Neil Fuller [Wed, 18 Feb 2015 09:42:31 +0000 (09:42 +0000)]
Merge "Update tzdata to tzdata2015a"

9 years agoImplement _FILE_OFFSET_BITS (mostly).
Elliott Hughes [Sat, 7 Feb 2015 06:28:49 +0000 (22:28 -0800)]
Implement _FILE_OFFSET_BITS (mostly).

I still don't think we can make stdio's fseeko and ftello work, but we can
have everything else, and very few programs use fseeko/ftello (and they can
just refrain from using _FILE_OFFSET_BITS and be no worse off than they are
today).

Bug: 11865851
Change-Id: Ic3cb409aae6713f4b345de954bcc4241fcd969ec

9 years agoMerge "Update NDK compatlib for 4.1.2."
Dan Albert [Wed, 18 Feb 2015 01:05:48 +0000 (01:05 +0000)]
Merge "Update NDK compatlib for 4.1.2."

9 years agoMerge "Update the NDK compatlib for 4.4.4 and ToT."
Dan Albert [Wed, 18 Feb 2015 01:05:43 +0000 (01:05 +0000)]
Merge "Update the NDK compatlib for 4.4.4 and ToT."

9 years agoUpdate NDK compatlib for 4.1.2.
Dan Albert [Fri, 13 Feb 2015 00:35:09 +0000 (16:35 -0800)]
Update NDK compatlib for 4.1.2.

The replacement new failures present in newer versions are present
here as well, with the following new issues:

XPASS std/localization/locale.categories/category.numeric/locale.nm.put/facet.num.put.members/put_long_double.pass.cpp

This is from the -NaN formatting fix in bionic. We previously had this
wrong, and the upstream test is also wrong. There's currently an XFAIL
for Android in this test because I haven't fixed the upstream test
yet. After that is done, I'll need to teach the test runner how to
XFAIL older Android versions...

FAIL std/localization/locale.categories/category.ctype/facet.ctype.special/facet.ctype.char.dtor/dtor.pass.cpp
    dtor.pass.cpp:39: int main(): assertion "globalMemCounter.checkDeleteArrayCalledEq(1)" failed

Haven't investigated this one yet. http://b/19412688

Note that this also needs the libgcc link ordering to be fixed in the
build system, as we'll otherwise depend on libgcc symbols from libc
that may or may not have been there.

The build fix can't be submitted because the proper link order causes
the libgcc unwinder to be used instead of the EHABI one:
http://b/18471342

Bug: 18471532
Change-Id: Icf560485a9b8f5ebbe01e4458703e62ec94df5e1

9 years agoUpdate the NDK compatlib for 4.4.4 and ToT.
Dan Albert [Tue, 10 Feb 2015 01:45:22 +0000 (17:45 -0800)]
Update the NDK compatlib for 4.4.4 and ToT.

Now passes all libc++ tests for these targets, with the exception of
the usual failing replacement new tests since libc uses new/delete for
things. I don't know if we can ever really fix these.

Bug: 18471532
Change-Id: Ibc0a15f26b0e4613249b5e15ecf3cf80e523467c

9 years agoMerge "Revert "Move use of __warnattr to __deprecated.""
Yabin Cui [Tue, 17 Feb 2015 23:56:29 +0000 (23:56 +0000)]
Merge "Revert "Move use of __warnattr to __deprecated.""

9 years agoRevert "Move use of __warnattr to __deprecated."
Yabin Cui [Tue, 17 Feb 2015 23:54:40 +0000 (23:54 +0000)]
Revert "Move use of __warnattr to __deprecated."

This reverts commit 4645c4f62df49d1ece5aabb7c9fa44ca34d30272.

Change-Id: I6c1062d54a4e2c4f41ce7a403e4e7840e6339146

9 years agoMerge "Move use of __warnattr to __deprecated."
Yabin Cui [Tue, 17 Feb 2015 23:10:30 +0000 (23:10 +0000)]
Merge "Move use of __warnattr to __deprecated."

9 years agoMove use of __warnattr to __deprecated.
Yabin Cui [Tue, 17 Feb 2015 20:39:39 +0000 (12:39 -0800)]
Move use of __warnattr to __deprecated.

clang don't support warning attribute. Replacing warning attriubte with
deprecated attribute can achieve the same behavior whether compiled by
gcc or clang.

Bug: 19340053
Change-Id: I064432b81cf55212458edbc749eb72dc15a810fb

9 years agoMerge "Accuracy tests for libm"
Christopher Ferris [Tue, 17 Feb 2015 21:41:41 +0000 (21:41 +0000)]
Merge "Accuracy tests for libm"

9 years agoMerge "Change getgrent error declaration message to proper style."
Yabin Cui [Tue, 17 Feb 2015 21:29:21 +0000 (21:29 +0000)]
Merge "Change getgrent error declaration message to proper style."

9 years agoMerge "Fix sysconf for _SC_CHILD_MAX and _SC_OPEN_MAX."
Elliott Hughes [Tue, 17 Feb 2015 21:20:39 +0000 (21:20 +0000)]
Merge "Fix sysconf for _SC_CHILD_MAX and _SC_OPEN_MAX."

9 years agoChange getgrent error declaration message to proper style.
Yabin Cui [Tue, 17 Feb 2015 19:33:42 +0000 (11:33 -0800)]
Change getgrent error declaration message to proper style.

Bug: 19340053
Change-Id: I9c47a2016f356d171a5f2082acb8391d81e019b2

9 years agoFix sysconf for _SC_CHILD_MAX and _SC_OPEN_MAX.
Elliott Hughes [Tue, 17 Feb 2015 17:55:58 +0000 (09:55 -0800)]
Fix sysconf for _SC_CHILD_MAX and _SC_OPEN_MAX.

Change-Id: I656f613166bd604f35b31e5ec042a5230c6b82b8

9 years agoMerge "Fix memchr overflow."
Elliott Hughes [Tue, 17 Feb 2015 17:54:39 +0000 (17:54 +0000)]
Merge "Fix memchr overflow."

9 years agoMerge "POSIX says flock is in <sys/file.h>, not <unistd.h>."
Elliott Hughes [Tue, 17 Feb 2015 17:09:34 +0000 (17:09 +0000)]
Merge "POSIX says flock is in <sys/file.h>, not <unistd.h>."

9 years agoPOSIX says flock is in <sys/file.h>, not <unistd.h>.
Elliott Hughes [Mon, 16 Feb 2015 20:19:46 +0000 (12:19 -0800)]
POSIX says flock is in <sys/file.h>, not <unistd.h>.

Bug: 19392265
Change-Id: I14d0b56883f0354e13db04a1d140b8f60dae08d7

9 years agoUpdate tzdata to tzdata2015a
Neil Fuller [Mon, 2 Feb 2015 11:24:22 +0000 (11:24 +0000)]
Update tzdata to tzdata2015a

Time Zone Data v. 2015a (Released 2015-01-29)
http://www.iana.org/time-zones/repository/releases/tzdata2015a.tar.gz

Information from NEWS:

Release 2015a - 2015-01-29 22:35:20 -0800

  Changes affecting future time stamps

    The Mexican state of Quintana Roo, represented by America/Cancun,
    will shift from Central Time with DST to Eastern Time without DST
    on 2015-02-01 at 02:00.  (Thanks to Steffen Thorsen and Gwillim Law.)

    Chile will not change clocks in April or thereafter; its new standard time
    will be its old daylight saving time.  This affects America/Santiago,
    Pacific/Easter, and Antarctica/Palmer.  (Thanks to Juan Correa.)

    New leap second 2015-06-30 23:59:60 UTC as per IERS Bulletin C 49.
    (Thanks to Tim Parenti.)

  Changes affecting past time stamps

    Iceland observed DST in 1919 and 1921, and its 1939 fallback
    transition was Oct. 29, not Nov. 29.  Remove incorrect data from
    Shanks about time in Iceland between 1837 and 1908.

    Some more zones have been turned into links, when they differed
    from existing zones only for older time stamps.  As usual,
    these changes affect UTC offsets in pre-1970 time stamps only.
    Their old contents have been moved to the 'backzone' file.
    The affected zones are: Asia/Aden, Asia/Bahrain, Asia/Kuwait,
    and Asia/Muscat.

Bug: 19212588
(cherry-picked from commit 700eb048fb6df8805245097d73a87384108fdf67)

Change-Id: I23d94982f634889ab9dcef28cf4a8853224bfd9c

9 years agoMerge "Regression test for NDK bug 80199."
Elliott Hughes [Sun, 15 Feb 2015 00:18:22 +0000 (00:18 +0000)]
Merge "Regression test for NDK bug 80199."

9 years agoRegression test for NDK bug 80199.
Elliott Hughes [Sat, 14 Feb 2015 22:11:50 +0000 (14:11 -0800)]
Regression test for NDK bug 80199.

Bionic never had this bug, but since the proposed fix is to remove the NDK's
broken code, we should add a regression test here.

Bug: https://code.google.com/p/android/issues/detail?id=80199
Change-Id: I4de21b5da9913cef990bc4d05a7e27562a71a02b

9 years agoFix memchr overflow.
Elliott Hughes [Sat, 14 Feb 2015 21:21:22 +0000 (13:21 -0800)]
Fix memchr overflow.

The overflow's actually in the generic C implementation of memchr.

While I'm here, let's switch our generic memrchr to the OpenBSD version too.

Bug: https://code.google.com/p/android/issues/detail?id=147048
Change-Id: I296ae06a1ee196d2c77c95a22f11ee4d658962da

9 years agoMerge "Declare getgrent/setgrent/endgrent as missing."
Yabin Cui [Sat, 14 Feb 2015 02:37:19 +0000 (02:37 +0000)]
Merge "Declare getgrent/setgrent/endgrent as missing."

9 years agoMerge changes I88827aa0,Ib0b0987a
Dmitriy Ivanov [Sat, 14 Feb 2015 00:38:03 +0000 (00:38 +0000)]
Merge changes I88827aa0,Ib0b0987a

* changes:
  Fix: DT_DEBUG was acting as DT_REL on mips64
  Add missing SHT_LOOS/SHT_HIOS values

9 years agoFix: DT_DEBUG was acting as DT_REL on mips64
Dmitriy Ivanov [Sat, 14 Feb 2015 00:29:50 +0000 (16:29 -0800)]
Fix: DT_DEBUG was acting as DT_REL on mips64

Change-Id: I88827aa07d75d06d606c987e6270fcca3ae6216f

9 years agoAdd missing SHT_LOOS/SHT_HIOS values
Dmitriy Ivanov [Sat, 14 Feb 2015 00:25:43 +0000 (16:25 -0800)]
Add missing SHT_LOOS/SHT_HIOS values

Change-Id: Ib0b0987a7e85af7863c6ef894263b5980e32344d

9 years agoDeclare getgrent/setgrent/endgrent as missing.
Yabin Cui [Fri, 13 Feb 2015 18:52:35 +0000 (10:52 -0800)]
Declare getgrent/setgrent/endgrent as missing.

Bug: 19340053
Change-Id: I42bfeda95e6f262e2e74ab47336ea346c2de7e4a

9 years agoMerge "Make .note.android.ident section type SH_NOTE"
Colin Cross [Fri, 13 Feb 2015 18:38:02 +0000 (18:38 +0000)]
Merge "Make .note.android.ident section type SH_NOTE"

9 years agoAccuracy tests for libm
Jingwei Zhang [Tue, 2 Sep 2014 13:39:14 +0000 (21:39 +0800)]
Accuracy tests for libm

This patch adds more tests for math functions to address coverage
issue of math functions discussed in:
https://android-review.googlesource.com/#/c/49653/
https://android-review.googlesource.com/#/c/94780/

These are data sets used in regression tests for the Intel the math library (libm). They were collected over a long period of testing various libm implementations.
The data sets contain function specific data (special and corner cases such as +/-0, maximum/minimum normalized numbers, +/-infinity, QNaN/SNaN, maximum/minimum denormal numbers, arguments that would produce close to overflow/underflow results, known hard-to-round cases, etc), implementation specific data (arguments close to table look-up values for different polynomial approximations, worst cases for range reduction algorithms) and other data with interesting bit patterns.
The reference values are computed with Maple and were converted into hexadecimal format.

Change-Id: I7177c282937369eae98f25d02134e4fc3beadde8
Signed-off-by: Jingwei Zhang <jingwei.zhang@intel.com>
Signed-off-by: Mingwei Shi <mingwei.shi@intel.com>
9 years agoMake .note.android.ident section type SH_NOTE
Colin Cross [Fri, 13 Feb 2015 01:03:22 +0000 (17:03 -0800)]
Make .note.android.ident section type SH_NOTE

The .note.android.ident section is only used by GDB, which doesn't
care what section type the section is, but it would be convenient
for readelf -n to be able to find the section too.

The old way of getting the .note.android.ident section to be of type
SH_NOTE involved compiling from .c to .s using gcc, running sed to
change progbits to note, and then compiling from .s to .o using gcc.
Since crtbrand.c only contains a section containing data, a
crtbrand.S can be checked in that will compile on all platforms,
avoiding the need for sed.

Also add crtbrand.o to crtbegin_so.o so that libraries also get
the note, and to the crt workaround in arm libc.so.

Change-Id: Ica71942a6af4553b56978ceaa288b3f4c15ebfa2

9 years agoMerge "bionic: benchmarks: Add BM_semaphore_sem_post_sem_wait"
Mark Salyzyn [Thu, 12 Feb 2015 20:52:57 +0000 (20:52 +0000)]
Merge "bionic: benchmarks: Add BM_semaphore_sem_post_sem_wait"

9 years agoMerge "Remove no-op sed step when compiling crtbrand.o"
Colin Cross [Thu, 12 Feb 2015 18:16:47 +0000 (18:16 +0000)]
Merge "Remove no-op sed step when compiling crtbrand.o"

9 years agobionic: benchmarks: Add BM_semaphore_sem_post_sem_wait
Mark Salyzyn [Tue, 10 Feb 2015 16:34:35 +0000 (08:34 -0800)]
bionic: benchmarks: Add BM_semaphore_sem_post_sem_wait

This test reports the overhead of sem_post to sem_wake for a low thread count
and a high thread count.

Change-Id: Ic30dcc8a78d754979117446bf3a28b7575cabac7

9 years agoRemove no-op sed step when compiling crtbrand.o
Colin Cross [Thu, 12 Feb 2015 01:40:45 +0000 (17:40 -0800)]
Remove no-op sed step when compiling crtbrand.o

crtbrand.c was compiled to a .s file, run through a sed script
to translate a %progbits to %note, and the compiled to .o.
However, when the sed command was copied from the original source
it was not updated to use the new name of the section (.note.ABI-tag
to .note.android.ident), so it didn't modify the file.  Since the
section has been generated with type %progbits instead of %note for
two years, just delete the whole sed step.

Change-Id: Id78582e9b43b628afec4eed22a088283132f0742

9 years ago[MIPS] Fix setjmp signals
Duane Sand [Mon, 26 Jan 2015 22:48:48 +0000 (14:48 -0800)]
[MIPS] Fix setjmp signals

Include full 16-byte Mips sigset_t signal mask within jump buffer.
Call sigprocmask instead of sigblockmask/sigsetmask to get/set full signal mask.
Include sigsetjmp's savesigs arg inside jmp_buf, instead of following it.
Reserve room for future extensions.
Preserve historically-large mips32 _JBLEN size.
Eliminate redundancy: code setjmp and _setjmp as tail calls into sigsetjmp,
and make longjmp and _longjmp aliases of siglongjmp.

Change-Id: Ie79137cf059228c1a51344ebb20d3a9a40b4a252

9 years agoMerge "[MIPS] Fix Mips64 build of libm"
Christopher Ferris [Wed, 11 Feb 2015 21:07:46 +0000 (21:07 +0000)]
Merge "[MIPS] Fix Mips64 build of libm"

9 years ago[MIPS] Fix Mips64 build of libm
Duane Sand [Wed, 11 Feb 2015 20:51:42 +0000 (12:51 -0800)]
[MIPS] Fix Mips64 build of libm

Change-Id: Ib742b408d4f21d5ec8d455a3fea7c08165c7f321

9 years agoMerge "Fix our ftw tests."
Elliott Hughes [Tue, 10 Feb 2015 23:34:18 +0000 (23:34 +0000)]
Merge "Fix our ftw tests."

9 years agoFix our ftw tests.
Elliott Hughes [Tue, 10 Feb 2015 22:15:33 +0000 (14:15 -0800)]
Fix our ftw tests.

SELinux denies access to some files in /sys, so we can't just trawl
through that asserting general truths. Instead, create a small known
tree.

Sadly neither ftw nor nftw takes user callback data, otherwise it would
be nice to assert that we visit all the expected nodes.

Bug: 19252748
Change-Id: Ib5309c38aaef53e6030281191a265a8d5a619044

9 years agoMerge "Define MAXHOSTNAMELEN explicitly in source files."
Yabin Cui [Tue, 10 Feb 2015 17:51:33 +0000 (17:51 +0000)]
Merge "Define MAXHOSTNAMELEN explicitly in source files."

9 years agoMerge "bionic: benchmarks: Add BM_semaphore_sem_post"
Mark Salyzyn [Tue, 10 Feb 2015 15:16:33 +0000 (15:16 +0000)]
Merge "bionic: benchmarks: Add BM_semaphore_sem_post"

9 years agoDefine MAXHOSTNAMELEN explicitly in source files.
Yabin Cui [Mon, 9 Feb 2015 21:58:28 +0000 (13:58 -0800)]
Define MAXHOSTNAMELEN explicitly in source files.

Bug: 19093777
Bug: 19092844
Change-Id: I0778507ca2d5c008abb9a6d6315d0909263a8817

9 years agobionic: benchmarks: Add BM_semaphore_sem_post
Mark Salyzyn [Mon, 9 Feb 2015 16:18:10 +0000 (08:18 -0800)]
bionic: benchmarks: Add BM_semaphore_sem_post

This test tries its best to report the producer side underlying futex
wake syscall overhead wthin sem_post. It does not measure the time it
takes for the wakeup to propagate to the consumer. It suffers from
clock_gettime syscall overhead, so subtract that. Lock the CPU speed
for consistent results as we may not reach >50% cpu utilization.

Change-Id: I02fa9dab2e6ac27202f0290115150bd3c8de00f2

9 years agoMerge "Cleanup the libm makefile."
Christopher Ferris [Tue, 10 Feb 2015 00:31:05 +0000 (00:31 +0000)]
Merge "Cleanup the libm makefile."

9 years agoMerge "Switch pthread_cond_t to <stdatomic.h>."
Yabin Cui [Tue, 10 Feb 2015 00:10:45 +0000 (00:10 +0000)]
Merge "Switch pthread_cond_t to <stdatomic.h>."

9 years agoSwitch pthread_cond_t to <stdatomic.h>.
Yabin Cui [Thu, 29 Jan 2015 20:13:33 +0000 (12:13 -0800)]
Switch pthread_cond_t to <stdatomic.h>.

Bug: 17574458
Change-Id: Ic7f79861df4fe751cfa6c6b20b71123cc31e7114

9 years agoCleanup the libm makefile.
Christopher Ferris [Fri, 23 Jan 2015 01:16:33 +0000 (17:16 -0800)]
Cleanup the libm makefile.

Change-Id: I3d95f75da24ee3b064a9cce67586b0e6f32272d8

9 years agoMerge "Stop including <linux/param.h> in <sys/param.h>."
Yabin Cui [Mon, 9 Feb 2015 19:46:45 +0000 (19:46 +0000)]
Merge "Stop including <linux/param.h> in <sys/param.h>."

9 years agoMerge "Remove stray log line."
Narayan Kamath [Mon, 9 Feb 2015 18:57:50 +0000 (18:57 +0000)]
Merge "Remove stray log line."

9 years agoRemove stray log line.
Narayan Kamath [Mon, 9 Feb 2015 17:59:22 +0000 (17:59 +0000)]
Remove stray log line.

Change-Id: Id948ae37a91dadd9d40e91de97f18774bdccfc78

9 years agoMerge "Display errors instead of running infinitely."
Christopher Ferris [Sat, 7 Feb 2015 02:18:39 +0000 (02:18 +0000)]
Merge "Display errors instead of running infinitely."

9 years agoDisplay errors instead of running infinitely.
Christopher Ferris [Tue, 3 Feb 2015 23:16:03 +0000 (15:16 -0800)]
Display errors instead of running infinitely.

When there is an error detected, the code runs forever and then times
out without any indication of what happened. Change it so that error
messages are printed and the test fails.

Change-Id: Id3160fc2f394984de0157356594fd8b40de66b4a

9 years agoMerge "Switch kernel header parsing to python libclang"
Tao Bao [Fri, 6 Feb 2015 23:25:29 +0000 (23:25 +0000)]
Merge "Switch kernel header parsing to python libclang"

9 years agoSwitch kernel header parsing to python libclang
Tao Bao [Wed, 28 Jan 2015 18:07:51 +0000 (10:07 -0800)]
Switch kernel header parsing to python libclang

Replace the tokenizer in cpp.py with libclang.

Bug: 18937958
Change-Id: I27630904c6d2849418cd5ca3d3c612ec3078686d

9 years agoMerge "Use LOCAL_C_INCLUDES instead of LOCAL_CFLAGS for include dirs"
Colin Cross [Fri, 6 Feb 2015 22:45:59 +0000 (22:45 +0000)]
Merge "Use LOCAL_C_INCLUDES instead of LOCAL_CFLAGS for include dirs"

9 years agoUse LOCAL_C_INCLUDES instead of LOCAL_CFLAGS for include dirs
Colin Cross [Fri, 6 Feb 2015 22:16:17 +0000 (14:16 -0800)]
Use LOCAL_C_INCLUDES instead of LOCAL_CFLAGS for include dirs

Change-Id: If61d6faae5d91c3bbe198e733e5922cd877ba353

9 years agoMerge "Add path separator in bionic-unit-tests program path."
Yabin Cui [Thu, 5 Feb 2015 21:55:18 +0000 (21:55 +0000)]
Merge "Add path separator in bionic-unit-tests program path."

9 years agoMerge "Use the libcxxabi __cxa_demangle."
Elliott Hughes [Thu, 5 Feb 2015 20:21:35 +0000 (20:21 +0000)]
Merge "Use the libcxxabi __cxa_demangle."

9 years agoMerge "Apparently, logd does need the NULs."
Elliott Hughes [Thu, 5 Feb 2015 20:05:45 +0000 (20:05 +0000)]
Merge "Apparently, logd does need the NULs."

9 years agoUse the libcxxabi __cxa_demangle.
Elliott Hughes [Thu, 5 Feb 2015 20:05:34 +0000 (12:05 -0800)]
Use the libcxxabi __cxa_demangle.

Bug: 18930951
Change-Id: If2ea3633ac6dfd4ec975c0f16bdf3dd58e05c614

9 years agoApparently, logd does need the NULs.
Elliott Hughes [Thu, 5 Feb 2015 20:02:04 +0000 (12:02 -0800)]
Apparently, logd does need the NULs.

Change-Id: I04b834e65c26c5821b952f78a0de7f92527cbdba

9 years agoAdd path separator in bionic-unit-tests program path.
Yabin Cui [Thu, 5 Feb 2015 04:53:56 +0000 (20:53 -0800)]
Add path separator in bionic-unit-tests program path.

Bug: 19220800
Change-Id: I4245b1270363544cd5f083fe7ea7d9b11b46cdc0

9 years agoMerge "Add signal handling in bionic gtest main."
Yabin Cui [Thu, 5 Feb 2015 04:21:03 +0000 (04:21 +0000)]
Merge "Add signal handling in bionic gtest main."

9 years agoAdd signal handling in bionic gtest main.
Yabin Cui [Tue, 3 Feb 2015 03:14:05 +0000 (19:14 -0800)]
Add signal handling in bionic gtest main.

The two bugs are very closely related and code amount is very small,
So I think they may be fixed in one change.

Bug: 19128558
Bug: 19129994
Change-Id: I44a35398e64dfca7e9676428cb8f4026e8f6e488

9 years agoMerge "linker: Use load_bias to adjust search address in elf_addr_lookup"
Dimitry Ivanov [Thu, 5 Feb 2015 02:23:34 +0000 (02:23 +0000)]
Merge "linker: Use load_bias to adjust search address in elf_addr_lookup"

9 years agoMerge changes Ib87855e8,I4b46ae20,I66364a5c,Id3fcf680
Colin Cross [Thu, 5 Feb 2015 01:57:27 +0000 (01:57 +0000)]
Merge changes Ib87855e8,I4b46ae20,I66364a5c,Id3fcf680

* changes:
  Use LOCAL_LDFLAGS_64 instead of enumerating 64-bit architectures
  Fix typo in cpu variant makefile depenendency for arm64
  Remove libc_static_common_src_files
  Share LP32 makefile settings between arches

9 years agoUse LOCAL_LDFLAGS_64 instead of enumerating 64-bit architectures
Colin Cross [Thu, 5 Feb 2015 01:40:29 +0000 (17:40 -0800)]
Use LOCAL_LDFLAGS_64 instead of enumerating 64-bit architectures

LOCAL_LDFLAGS_<arch> was being set for arm64, mips64, and x86_64.
Use LOCAL_LDFLAGS_64 instead.

Change-Id: Ib87855e8a7783f82461b707fffa7f1663e6be9c8

9 years agoFix typo in cpu variant makefile depenendency for arm64
Colin Cross [Thu, 5 Feb 2015 01:38:27 +0000 (17:38 -0800)]
Fix typo in cpu variant makefile depenendency for arm64

cpu_variank_mk -> cpu_variant_mk

Change-Id: I4b46ae20de72ed2dc8820d62ac35ddc6b915a195

9 years agoRemove libc_static_common_src_files
Colin Cross [Thu, 5 Feb 2015 01:37:23 +0000 (17:37 -0800)]
Remove libc_static_common_src_files

libc_static_common_src_files is never set after
c3f114037dbf028896310609fd28cf2b3da99c4d, remove the remaining
references to it.

Change-Id: I66364a5c1b031ad69d608f6f44244049192944f6

9 years agoShare LP32 makefile settings between arches
Colin Cross [Thu, 5 Feb 2015 01:35:49 +0000 (17:35 -0800)]
Share LP32 makefile settings between arches

Add <var>_32 to patch-up-arch-specific-flags, and move the LP32
cruft varaibles from the 32-bit arch specific makefiles into the
top level Android.mk.

Change-Id: Id3fcf6805d4af048c2524c94b1295416ebe7d057

9 years agolinker: Use load_bias to adjust search address in elf_addr_lookup
Chris Dearman [Thu, 14 Nov 2013 01:22:33 +0000 (17:22 -0800)]
linker: Use load_bias to adjust search address in elf_addr_lookup

load_bias is the delta between a symbols address in memory and the value
contained in the symbol table.

Change-Id: I35c87b69025b64f7e4a1418ca766618f4533fb3e

9 years agoMerge "Fix poll/ppoll fortify test to avoid hanging in failed fortify clang test."
Yabin Cui [Thu, 5 Feb 2015 01:14:32 +0000 (01:14 +0000)]
Merge "Fix poll/ppoll fortify test to avoid hanging in failed fortify clang test."

9 years agoStop including <linux/param.h> in <sys/param.h>.
Yabin Cui [Wed, 4 Feb 2015 02:40:44 +0000 (18:40 -0800)]
Stop including <linux/param.h> in <sys/param.h>.

Bug: 19092844
Change-Id: I1a4a0c84a196a187ea32e58491de744cf3c0f404