OSDN Git Service

android-x86/bionic.git
10 years agoFix wchar_t signedness problems found on x86-64.
Elliott Hughes [Fri, 2 May 2014 02:03:18 +0000 (19:03 -0700)]
Fix wchar_t signedness problems found on x86-64.

The existing tests caught this.

Change-Id: I6269844ae4301fd2c596241a59e97eb67ef166fa

10 years agoMerge "Fix wchar.cpp MIPS narrowing conversion build failure."
Elliott Hughes [Fri, 2 May 2014 00:15:42 +0000 (00:15 +0000)]
Merge "Fix wchar.cpp MIPS narrowing conversion build failure."

10 years agoFix wchar.cpp MIPS narrowing conversion build failure.
Elliott Hughes [Fri, 2 May 2014 00:14:59 +0000 (17:14 -0700)]
Fix wchar.cpp MIPS narrowing conversion build failure.

Change-Id: Id9103c78958d60337dbdb807b11256c1b31c632a

10 years agoMerge "Fix wchar.cpp signed/unsigned comparison build failure."
Elliott Hughes [Thu, 1 May 2014 23:50:35 +0000 (23:50 +0000)]
Merge "Fix wchar.cpp signed/unsigned comparison build failure."

10 years agoFix wchar.cpp signed/unsigned comparison build failure.
Elliott Hughes [Thu, 1 May 2014 23:49:55 +0000 (16:49 -0700)]
Fix wchar.cpp signed/unsigned comparison build failure.

Change-Id: Id20b91f3d57c4430987b5cc88ac99c245801d73b

10 years agoMerge "Switch to a working UTF-8 mb/wc implementation."
Elliott Hughes [Thu, 1 May 2014 23:12:41 +0000 (23:12 +0000)]
Merge "Switch to a working UTF-8 mb/wc implementation."

10 years agoSwitch to a working UTF-8 mb/wc implementation.
Elliott Hughes [Thu, 1 May 2014 05:03:12 +0000 (22:03 -0700)]
Switch to a working UTF-8 mb/wc implementation.

Although glibc gets by with an 8-byte mbstate_t, OpenBSD uses 12 bytes (of
the 128 bytes it reserves!).

We can actually implement UTF-8 encoding/decoding with a 0-byte mbstate_t
which means we can make things work on LP32 too, as long as we accept the
limitation that the caller needs to present us with a complete sequence
before we'll process it.

Our behavior is fine when going from characters to bytes; we just
update the source wchar_t** to say how far through the input we got.

I'll come back and use the 4 bytes we do have to cope with byte sequences
split across multiple input buffers. The fact that we don't support
UTF-8 sequences longer than 4 bytes plus the fact that the first byte of
a UTF-8 sequence encodes the length means we shouldn't need the other
fields OpenBSD used (at the cost of some recomputation in cases where a
sequence is split across buffers).

This patch also makes the minimal changes necessary to setlocale(3) to
make us behave like glibc when an app requests UTF-8. (The difference
being that our "C" locale is the same as our "C.UTF-8" locale.)

Change-Id: Ied327a8c4643744b3611bf6bb005a9b389ba4c2f

10 years agoMerge "Add ability to run on host for x86/x86_64."
Christopher Ferris [Thu, 1 May 2014 21:03:27 +0000 (21:03 +0000)]
Merge "Add ability to run on host for x86/x86_64."

10 years agoAdd ability to run on host for x86/x86_64.
Christopher Ferris [Thu, 1 May 2014 20:44:57 +0000 (13:44 -0700)]
Add ability to run on host for x86/x86_64.

Change-Id: I063cb71ce0f7247cefacaa653c7b87b7b3e72025

10 years agoMerge "Modify hard-coded directory."
Christopher Ferris [Thu, 1 May 2014 20:19:20 +0000 (20:19 +0000)]
Merge "Modify hard-coded directory."

10 years agoModify hard-coded directory.
Christopher Ferris [Thu, 1 May 2014 20:00:32 +0000 (13:00 -0700)]
Modify hard-coded directory.

Use the ANDROID_DATA environment variable instead of the hard-coded
directory for these benchmarks.

Change-Id: I00bae7b4a24e81e77fc8f52e1fe99f4d4918f520

10 years agoMerge "Keep the kernel header scrubber's data structures in sync."
Elliott Hughes [Thu, 1 May 2014 17:28:12 +0000 (17:28 +0000)]
Merge "Keep the kernel header scrubber's data structures in sync."

10 years agoKeep the kernel header scrubber's data structures in sync.
Elliott Hughes [Thu, 1 May 2014 17:17:27 +0000 (10:17 -0700)]
Keep the kernel header scrubber's data structures in sync.

If you rewrite the tokens of a #if you need to rewrite the expression to match
because either might be used later. This was showing up as SIGRTMAX being
rewritten in a #define but not in the #ifndef that guarded it, for which case
I've added a unit test.

Change-Id: I6929675461a1afe272edd667594529fd84a3dc4d

10 years agoMerge "Make SIGRTMIN hide the real-time signals we use internally."
Elliott Hughes [Wed, 30 Apr 2014 18:30:14 +0000 (18:30 +0000)]
Merge "Make SIGRTMIN hide the real-time signals we use internally."

10 years agoMerge "Fix the CPU_*_S tests."
Elliott Hughes [Wed, 30 Apr 2014 18:29:31 +0000 (18:29 +0000)]
Merge "Fix the CPU_*_S tests."

10 years agoFix the CPU_*_S tests.
Elliott Hughes [Wed, 30 Apr 2014 17:45:35 +0000 (10:45 -0700)]
Fix the CPU_*_S tests.

Change-Id: Id67c48b9c12a20b01309490670438bfcd9163465

10 years agoMake SIGRTMIN hide the real-time signals we use internally.
Elliott Hughes [Wed, 30 Apr 2014 16:45:40 +0000 (09:45 -0700)]
Make SIGRTMIN hide the real-time signals we use internally.

__SIGRTMIN will continue to tell the truth. This matches glibc's
behavior (as evidenced by the fact that we don't need a special case
in the strsignal test now).

Change-Id: I1abe1681d516577afa8cd39c837ef12467f68dd2

10 years agoMerge "Fix cpu_set_t"
Calin Juravle [Wed, 30 Apr 2014 16:09:07 +0000 (16:09 +0000)]
Merge "Fix cpu_set_t"

10 years agoMerge "Adds quick_exit(3) and at_quick_exit(3) from freebsd"
Dan Albert [Wed, 30 Apr 2014 15:39:13 +0000 (15:39 +0000)]
Merge "Adds quick_exit(3) and at_quick_exit(3) from freebsd"

10 years agoFix cpu_set_t
Calin Juravle [Tue, 29 Apr 2014 19:25:26 +0000 (20:25 +0100)]
Fix cpu_set_t

- extend CPU_SETSIZE for LP64
- fix CPU_(AND|OR|XOR) macros
- fix CPU_OP_S macro
- fix __sched_cpucount
- tidy up the code

Change-Id: I741afff4d0c473e8a1ee6b4141981dc24467e0d4

10 years agoMerge "Reserve space in mbstate to allow for proper wchar support"
Calin Juravle [Wed, 30 Apr 2014 10:34:08 +0000 (10:34 +0000)]
Merge "Reserve space in mbstate to allow for proper wchar support"

10 years agoMerge "Switch to OpenBSD stdio wide printf functions."
Elliott Hughes [Wed, 30 Apr 2014 03:16:40 +0000 (03:16 +0000)]
Merge "Switch to OpenBSD stdio wide printf functions."

10 years agoAdds quick_exit(3) and at_quick_exit(3) from freebsd
Dan Albert [Wed, 30 Apr 2014 00:49:06 +0000 (17:49 -0700)]
Adds quick_exit(3) and at_quick_exit(3) from freebsd

Change-Id: I4fe88abd8f7b8aa45e58aeb2529d59a8d555d338

10 years agoSwitch to OpenBSD stdio wide printf functions.
Elliott Hughes [Wed, 30 Apr 2014 00:39:29 +0000 (17:39 -0700)]
Switch to OpenBSD stdio wide printf functions.

Change-Id: Icf4f8685d021ec6b7482ca1cc021ce8184098e4a

10 years agoMerge "Switch to OpenBSD stdio wide get/put functions."
Elliott Hughes [Wed, 30 Apr 2014 00:38:32 +0000 (00:38 +0000)]
Merge "Switch to OpenBSD stdio wide get/put functions."

10 years agoSwitch to OpenBSD stdio wide get/put functions.
Elliott Hughes [Wed, 30 Apr 2014 00:08:03 +0000 (17:08 -0700)]
Switch to OpenBSD stdio wide get/put functions.

Change-Id: I71f8769cdea874e55d397ca7682d9d4e659d3dcb

10 years agoMerge "Switch to the OpenBSD implementations of the wide scanf functions."
Elliott Hughes [Tue, 29 Apr 2014 23:57:11 +0000 (23:57 +0000)]
Merge "Switch to the OpenBSD implementations of the wide scanf functions."

10 years agoSwitch to the OpenBSD implementations of the wide scanf functions.
Elliott Hughes [Tue, 29 Apr 2014 23:28:56 +0000 (16:28 -0700)]
Switch to the OpenBSD implementations of the wide scanf functions.

This also gets us the C99 wcstoimax and wcstoumax, and a working fgetwc and
ungetwc, all of which are needed in the implementation.

This also brings several other files closer to upstream.

Change-Id: I23b025a8237a6dbb9aa50d2a96765ea729a85579

10 years agoMerge "Fix build (signed char issue)."
Elliott Hughes [Tue, 29 Apr 2014 23:06:37 +0000 (23:06 +0000)]
Merge "Fix build (signed char issue)."

10 years agoFix build (signed char issue).
Elliott Hughes [Tue, 29 Apr 2014 23:05:58 +0000 (16:05 -0700)]
Fix build (signed char issue).

Change-Id: I05d78f4c1599ed9a0c1285f9eb1e89bc2f55c24d

10 years agoMerge "Switch to the OpenBSD wcsto* functions."
Elliott Hughes [Tue, 29 Apr 2014 22:33:15 +0000 (22:33 +0000)]
Merge "Switch to the OpenBSD wcsto* functions."

10 years agoSwitch to the OpenBSD wcsto* functions.
Elliott Hughes [Tue, 29 Apr 2014 21:46:56 +0000 (14:46 -0700)]
Switch to the OpenBSD wcsto* functions.

This replaces a partial set of non-functional functions with a complete
set of functions, all of which actually work.

This requires us to implement mbsnrtowcs and wcsnrtombs which completes
the set of what we need for libc++.

The mbsnrtowcs is basically a copy & paste of wcsnrtombs, but I'm going
to go straight to looking at using the OpenBSD UTF-8 implementation rather
than keep polishing our home-grown turd.

(This patch also opportunistically switches us over to upstream btowc,
mbrlen, and wctob, since they're all trivially expressed in terms of
other functions.)

Change-Id: I0f81443840de0f1aa73b96f0b51988976793a323

10 years agoReserve space in mbstate to allow for proper wchar support
Calin Juravle [Tue, 29 Apr 2014 14:48:34 +0000 (15:48 +0100)]
Reserve space in mbstate to allow for proper wchar support

Bug: 14382788
Change-Id: If023ac9bb65f95135cae7ebe89147e3985a69a96

10 years agoMerge "Fix potential makefile bug."
Christopher Ferris [Tue, 29 Apr 2014 15:56:06 +0000 (15:56 +0000)]
Merge "Fix potential makefile bug."

10 years agoFix potential makefile bug.
Christopher Ferris [Tue, 29 Apr 2014 04:10:37 +0000 (21:10 -0700)]
Fix potential makefile bug.

Since multilib is not set every time, it needs to be per module or
there is a change that another target will use the multilib value set
previously.

Change-Id: I5c30e18d5111705cb3f6e3d4cd9ef8a28c9b746c

10 years agoMerge "Add mbtowc and fix mbrtowc."
Elliott Hughes [Tue, 29 Apr 2014 00:54:43 +0000 (00:54 +0000)]
Merge "Add mbtowc and fix mbrtowc."

10 years agoAdd mbtowc and fix mbrtowc.
Elliott Hughes [Tue, 29 Apr 2014 00:51:13 +0000 (17:51 -0700)]
Add mbtowc and fix mbrtowc.

Change-Id: I48786cd82587e61188d40f6fd6e11ac05e857ae9

10 years agoMerge "Replace our broken wcswcs with the working upstream one."
Elliott Hughes [Mon, 28 Apr 2014 23:47:17 +0000 (23:47 +0000)]
Merge "Replace our broken wcswcs with the working upstream one."

10 years agoReplace our broken wcswcs with the working upstream one.
Elliott Hughes [Mon, 28 Apr 2014 23:28:51 +0000 (16:28 -0700)]
Replace our broken wcswcs with the working upstream one.

Change-Id: I2952684df5674d10f0564d92c2cd42597725c0e3

10 years agoMerge "Don't use so much stack in tzcode."
Elliott Hughes [Sat, 26 Apr 2014 04:51:55 +0000 (04:51 +0000)]
Merge "Don't use so much stack in tzcode."

10 years agoDon't use so much stack in tzcode.
Elliott Hughes [Fri, 25 Apr 2014 23:55:04 +0000 (16:55 -0700)]
Don't use so much stack in tzcode.

Bug: 14313703
Bug: https://code.google.com/p/android/issues/detail?id=61130
Change-Id: Id9b240fce20d4a6f2660792070e02b2b5f287655

10 years agoMerge "Pass si_code through to debuggerd."
Elliott Hughes [Sat, 26 Apr 2014 03:23:22 +0000 (03:23 +0000)]
Merge "Pass si_code through to debuggerd."

10 years agoPass si_code through to debuggerd.
Elliott Hughes [Fri, 25 Apr 2014 23:02:00 +0000 (16:02 -0700)]
Pass si_code through to debuggerd.

Because we re-raise various signals, we corrupt the si_code that debuggerd
sees when it ptraces our siginfo. One possible solution (shown here) is to
pass the original si_code value in the message we send to debuggerd.

Change-Id: I76f9aa2c0442e5cab611d132532409e700383907

10 years agoMerge "Fix brk/sbrk error checking."
Elliott Hughes [Sat, 26 Apr 2014 02:39:09 +0000 (02:39 +0000)]
Merge "Fix brk/sbrk error checking."

10 years agoFix brk/sbrk error checking.
Elliott Hughes [Sat, 26 Apr 2014 01:27:38 +0000 (18:27 -0700)]
Fix brk/sbrk error checking.

Note that the kernel returns the current break on error or if the requested
break is smaller than the minimum break, or the new break. I don't know where
we got the idea that the kernel could return -1.

Also optimizes the query case.

Also hides an accidentally-exported symbol for LP64.

Change-Id: I0fd6b8b14ddf1ae82935c0c3fc610da5cc74932e

10 years agoMerge "Reserve space in pthread structures for future extensibility."
Colin Cross [Thu, 24 Apr 2014 22:40:50 +0000 (22:40 +0000)]
Merge "Reserve space in pthread structures for future extensibility."

10 years agoReserve space in pthread structures for future extensibility.
Calin Juravle [Wed, 23 Apr 2014 16:35:04 +0000 (17:35 +0100)]
Reserve space in pthread structures for future extensibility.

Make our structures equal in size with glibc structures. This should
give us plenty of space to implement any missing feature.

Bug: 13278744
Bug: 12875898
(cherry picked from commit cf83fd77caa3128ecb6ac935e342cb0ea258b9ee)

Change-Id: I76968d31024eb51bc73887687e5ac492eb02a27f

10 years agoMerge "Allow liblog to pass failure reasons to debuggerd."
Elliott Hughes [Thu, 24 Apr 2014 20:29:43 +0000 (20:29 +0000)]
Merge "Allow liblog to pass failure reasons to debuggerd."

10 years agoMerge "Revert "Reserve space in pthread structures for future extensibility.""
Greg Hackmann [Thu, 24 Apr 2014 18:12:27 +0000 (18:12 +0000)]
Merge "Revert "Reserve space in pthread structures for future extensibility.""

10 years agoRevert "Reserve space in pthread structures for future extensibility."
Greg Hackmann [Thu, 24 Apr 2014 18:11:23 +0000 (18:11 +0000)]
Revert "Reserve space in pthread structures for future extensibility."

This reverts commit cf83fd77caa3128ecb6ac935e342cb0ea258b9ee.

Change-Id: I4f69e86ffe1c267a9018eb489d909237534c6cf9

10 years agoMerge "Remove an unsed include dir"
Calin Juravle [Thu, 24 Apr 2014 16:40:50 +0000 (16:40 +0000)]
Merge "Remove an unsed include dir"

10 years agoRemove an unsed include dir
Calin Juravle [Thu, 24 Apr 2014 16:28:05 +0000 (17:28 +0100)]
Remove an unsed include dir

Change-Id: I05aa1e7e0639a5ec5576bf7646f35d2e5b157c09

10 years agoMerge "Removed non-existing include dir"
Calin Juravle [Thu, 24 Apr 2014 16:29:49 +0000 (16:29 +0000)]
Merge "Removed non-existing include dir"

10 years agoRemoved non-existing include dir
Calin Juravle [Thu, 24 Apr 2014 16:17:02 +0000 (17:17 +0100)]
Removed non-existing include dir

Change-Id: I3e00a8471e6d94f596a34bdfdb26a0caba23dd82

10 years agoMerge "Reserve space in pthread structures for future extensibility."
Calin Juravle [Thu, 24 Apr 2014 11:35:36 +0000 (11:35 +0000)]
Merge "Reserve space in pthread structures for future extensibility."

10 years agoMerge "Clean-up _fpmath and fake_long_doubles"
Calin Juravle [Thu, 24 Apr 2014 10:12:57 +0000 (10:12 +0000)]
Merge "Clean-up _fpmath and fake_long_doubles"

10 years agoMerge "Fix fallout from host GCC upgrade."
Elliott Hughes [Thu, 24 Apr 2014 06:06:19 +0000 (06:06 +0000)]
Merge "Fix fallout from host GCC upgrade."

10 years agoFix fallout from host GCC upgrade.
Elliott Hughes [Thu, 24 Apr 2014 06:02:43 +0000 (23:02 -0700)]
Fix fallout from host GCC upgrade.

I'll raise a bug for the FD_ISSET fortification; we should do better too.

Change-Id: Id2bf277890ad06b010dc952e270d746714c2bea7

10 years agoAllow liblog to pass failure reasons to debuggerd.
Elliott Hughes [Wed, 23 Apr 2014 21:52:49 +0000 (14:52 -0700)]
Allow liblog to pass failure reasons to debuggerd.

assert(3) already does this, but LOG_ALWAYS_FATAL and LOG_ALWAYS_FATAL_IF
have been missing out.

Change-Id: I1d6214c4f792fa0d4ba3c14eded3fc9c332bd3c5

10 years agoMerge "Remove two dead files."
Elliott Hughes [Wed, 23 Apr 2014 19:54:43 +0000 (19:54 +0000)]
Merge "Remove two dead files."

10 years agoRemove two dead files.
Elliott Hughes [Wed, 23 Apr 2014 19:51:37 +0000 (12:51 -0700)]
Remove two dead files.

I forgot to remove these in the change that stopped building them
(eae5902e73dc4381811e08fd2334bf4a9300a928).

Change-Id: I4665fa9f6b9b80c51778d5e82b49406fe56db9a6

10 years agoMerge "prctl: Adds the PR_SET_TIMERSLACK_PID option"
Ruchi Kandoi [Wed, 23 Apr 2014 19:13:28 +0000 (19:13 +0000)]
Merge "prctl: Adds the PR_SET_TIMERSLACK_PID option"

10 years agoprctl: Adds the PR_SET_TIMERSLACK_PID option
Ruchi Kandoi [Wed, 23 Apr 2014 02:00:45 +0000 (19:00 -0700)]
prctl: Adds the PR_SET_TIMERSLACK_PID option

Allows software to set timer clack for other thread.

Change-Id: Ifa9d2cc7844ec581356bce15434609242fd898a8
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
10 years agoClean-up _fpmath and fake_long_doubles
Calin Juravle [Thu, 17 Apr 2014 17:17:32 +0000 (18:17 +0100)]
Clean-up _fpmath and fake_long_doubles

- promoted IEEEld2bits to fpmath since most of the where the same for
diffrent archs
- removed _fpmath
- reinstated weak_references
- moved isfinite and isnormal to libc
- clean up fake_long_doubles
- clean up some useless ifdefs
- added missing nexttoward* tests

Bug: 14134235
Change-Id: I95639c4885653fe47fd7dc0570ee5bb3389bbc6b

10 years agoReserve space in pthread structures for future extensibility.
Calin Juravle [Wed, 23 Apr 2014 16:35:04 +0000 (17:35 +0100)]
Reserve space in pthread structures for future extensibility.

Make our structures equal in size with glibc structures. This should
give us plenty of space to implement any missing feature.

Bug: 13278744
Bug: 12875898
Change-Id: I980017fd5942411ebc5ac32b2673a10a20db68c8

10 years agoMerge "Fix the glibc tests again."
Elliott Hughes [Wed, 23 Apr 2014 02:24:41 +0000 (02:24 +0000)]
Merge "Fix the glibc tests again."

10 years agoFix the glibc tests again.
Elliott Hughes [Wed, 23 Apr 2014 02:21:32 +0000 (19:21 -0700)]
Fix the glibc tests again.

The glibc tests are just a regular host binary; they don't require
that you're targeting x86 or x86_64. They do seem to pick up the
suffix of the target though, even though they're always 32-bit.

Change-Id: I689ca2a4f8d7b397afa4df722b95b0d7ec904bf6

10 years agoMerge "Tests for environment variables utility functions"
Elliott Hughes [Wed, 23 Apr 2014 02:10:39 +0000 (02:10 +0000)]
Merge "Tests for environment variables utility functions"

10 years agoTests for environment variables utility functions
Grigoriy Kraynov [Tue, 17 Sep 2013 11:44:22 +0000 (15:44 +0400)]
Tests for environment variables utility functions

This is the first patch from the new set of tests for Bionic standard functions.

Change-Id: Ie568788a24832394e597ad33f44a5c71cb33b51f
Signed-off-by: Grigoriy Kraynov <grigoriy.kraynov@intel.com>
10 years agoMerge "Switch to the upstream OpenBSD getenv/putenv/setenv implementation."
Elliott Hughes [Wed, 23 Apr 2014 01:52:06 +0000 (01:52 +0000)]
Merge "Switch to the upstream OpenBSD getenv/putenv/setenv implementation."

10 years agoMerge "Remove strntoimax and strntoumax from the future."
Elliott Hughes [Wed, 23 Apr 2014 01:50:08 +0000 (01:50 +0000)]
Merge "Remove strntoimax and strntoumax from the future."

10 years agoMerge "Get ANDROID_DATA dir from enviroment"
Dmitriy Ivanov [Wed, 23 Apr 2014 01:48:55 +0000 (01:48 +0000)]
Merge "Get ANDROID_DATA dir from enviroment"

10 years agoGet ANDROID_DATA dir from enviroment
Dmitriy Ivanov [Wed, 23 Apr 2014 00:48:08 +0000 (17:48 -0700)]
Get ANDROID_DATA dir from enviroment

Change-Id: Ie3675bd27bbc779fc2140f95b930eadd14838753

10 years agoRemove strntoimax and strntoumax from the future.
Elliott Hughes [Wed, 23 Apr 2014 00:56:42 +0000 (17:56 -0700)]
Remove strntoimax and strntoumax from the future.

Where do these turds come from?

Change-Id: Id9ad2cc85c6128aa63b5d56ff2aa455bde39a5eb

10 years agoSwitch to the upstream OpenBSD getenv/putenv/setenv implementation.
Elliott Hughes [Wed, 23 Apr 2014 00:41:00 +0000 (17:41 -0700)]
Switch to the upstream OpenBSD getenv/putenv/setenv implementation.

This fixes all the bugs found by the new tests.

Change-Id: Id5a5f9f39a0620208bafa053f871a044725b4795

10 years agoMerge "Follow up on building tests for both platforms"
Dmitriy Ivanov [Tue, 22 Apr 2014 23:03:21 +0000 (23:03 +0000)]
Merge "Follow up on building tests for both platforms"

10 years agoMerge "LP64 shouldn't include the non-standard <time64.h> cruft."
Elliott Hughes [Tue, 22 Apr 2014 22:53:15 +0000 (22:53 +0000)]
Merge "LP64 shouldn't include the non-standard <time64.h> cruft."

10 years agoFollow up on building tests for both platforms
Dmitriy Ivanov [Tue, 22 Apr 2014 22:22:25 +0000 (15:22 -0700)]
Follow up on building tests for both platforms

this makes run-on-host work properly

Change-Id: Iaed93ca9d96359b64bbeff995280ecd70fc9cc60

10 years agoLP64 shouldn't include the non-standard <time64.h> cruft.
Elliott Hughes [Tue, 22 Apr 2014 20:55:58 +0000 (13:55 -0700)]
LP64 shouldn't include the non-standard <time64.h> cruft.

This patch includes just enough to keep external/chromium_org building
until they switch 64-bit Android over to using the regular non-Android code.

Change-Id: Iecaf274efa46ae18a42d5e3439c5aa4f909177c1

10 years agoMerge "Build 32/64 bit versions of tests/benchmarks."
Christopher Ferris [Tue, 22 Apr 2014 20:45:01 +0000 (20:45 +0000)]
Merge "Build 32/64 bit versions of tests/benchmarks."

10 years agoMerge "Use -fvisibility=hidden to build libc_gdtoa."
Elliott Hughes [Tue, 22 Apr 2014 20:42:36 +0000 (20:42 +0000)]
Merge "Use -fvisibility=hidden to build libc_gdtoa."

10 years agoUse -fvisibility=hidden to build libc_gdtoa.
Elliott Hughes [Tue, 22 Apr 2014 19:32:49 +0000 (12:32 -0700)]
Use -fvisibility=hidden to build libc_gdtoa.

Bug: 12177745
Change-Id: Ia4ac71e1a3a237e764c809cc591622f8952ed642

10 years agoBuild 32/64 bit versions of tests/benchmarks.
Christopher Ferris [Tue, 22 Apr 2014 17:42:12 +0000 (10:42 -0700)]
Build 32/64 bit versions of tests/benchmarks.

Change-Id: I4d146377aac6573a214dfaa5cab5cce8b10538b4

10 years agoMerge "Fix WCHAR_MAX, WCHAR_MIN, WINT_MAX, and WINT_MIN."
Elliott Hughes [Tue, 22 Apr 2014 17:00:33 +0000 (17:00 +0000)]
Merge "Fix WCHAR_MAX, WCHAR_MIN, WINT_MAX, and WINT_MIN."

10 years agoMerge "Fix arm64/x86 build."
Torne (Richard Coles) [Tue, 22 Apr 2014 11:01:22 +0000 (11:01 +0000)]
Merge "Fix arm64/x86 build."

10 years agoFix arm64/x86 build.
Torne (Richard Coles) [Tue, 22 Apr 2014 10:59:26 +0000 (11:59 +0100)]
Fix arm64/x86 build.

Add missing second param to soinfo_link_image in a section that's not
compiled on ARM.

Bug: 13005501
Change-Id: Id0ede8e03da4e05b25c0aeb24a840f868031d4e8

10 years agoMerge "Allow sharing the RELRO section via a file."
Torne (Richard Coles) [Tue, 22 Apr 2014 10:39:49 +0000 (10:39 +0000)]
Merge "Allow sharing the RELRO section via a file."

10 years agoMerge "Support loading libraries to a reserved address."
Torne (Richard Coles) [Tue, 22 Apr 2014 10:33:33 +0000 (10:33 +0000)]
Merge "Support loading libraries to a reserved address."

10 years agoMerge "Add android_dlopen_ext() interface to linker."
Torne (Richard Coles) [Tue, 22 Apr 2014 10:33:06 +0000 (10:33 +0000)]
Merge "Add android_dlopen_ext() interface to linker."

10 years agoFix WCHAR_MAX, WCHAR_MIN, WINT_MAX, and WINT_MIN.
Elliott Hughes [Tue, 22 Apr 2014 01:09:46 +0000 (18:09 -0700)]
Fix WCHAR_MAX, WCHAR_MIN, WINT_MAX, and WINT_MIN.

GCC tells us everything we need to know. clang does its usual half-assed job.

Change-Id: Id4d664529b10345274602768cd564d3df717e931

10 years agoMerge "Clean up and document our hacks for building BSD source."
Elliott Hughes [Tue, 22 Apr 2014 00:34:01 +0000 (00:34 +0000)]
Merge "Clean up and document our hacks for building BSD source."

10 years agoClean up and document our hacks for building BSD source.
Elliott Hughes [Tue, 22 Apr 2014 00:13:46 +0000 (17:13 -0700)]
Clean up and document our hacks for building BSD source.

Change-Id: Ic591e22fa5b363bb68376b9f25814c0e5bd83fbf

10 years agoMerge "clang is >= 3.4 now; remove a hack."
Elliott Hughes [Mon, 21 Apr 2014 18:26:22 +0000 (18:26 +0000)]
Merge "clang is >= 3.4 now; remove a hack."

10 years agoMerge "Remove unused #defines from "arith.h"."
Elliott Hughes [Mon, 21 Apr 2014 18:17:45 +0000 (18:17 +0000)]
Merge "Remove unused #defines from "arith.h"."

10 years agoclang is >= 3.4 now; remove a hack.
Elliott Hughes [Mon, 21 Apr 2014 17:15:14 +0000 (10:15 -0700)]
clang is >= 3.4 now; remove a hack.

Change-Id: I8d976ee0fd88eca9779c87b5d7e27f46b4b1b489

10 years agoRemove unused #defines from "arith.h".
Elliott Hughes [Mon, 21 Apr 2014 16:59:00 +0000 (09:59 -0700)]
Remove unused #defines from "arith.h".

These were defined in upstream's per-arch files, so I carried them across,
but they aren't actually used anywhere.

Change-Id: Ica9796201dcd2caa5c34070c0dd226fc2050ba47

10 years agoMerge "Make libc signal handler output more like debuggerd."
Elliott Hughes [Sat, 19 Apr 2014 02:02:04 +0000 (02:02 +0000)]
Merge "Make libc signal handler output more like debuggerd."

10 years agoMerge "Fix for libgcc compat generation script."
Dmitriy Ivanov [Sat, 19 Apr 2014 01:12:58 +0000 (01:12 +0000)]
Merge "Fix for libgcc compat generation script."

10 years agoMake libc signal handler output more like debuggerd.
Elliott Hughes [Sat, 19 Apr 2014 00:39:25 +0000 (17:39 -0700)]
Make libc signal handler output more like debuggerd.

This has been annoying me for a while, because it's often quite misleading.

Today, for example, I saw:

  Fatal signal 13 (SIGPIPE) at 0x6573 (code=0), thread 25971 (top)

where the apparent address is actually the pid of the signal source (in this
case the kernel on behalf of the thread itself).

This patch isn't as fancy as strace, but it at least means we never say
anything misleading. We could decode the si_code field like strace and
debuggerd, but I'm reluctant to do that without some way to share the code
between at least bionic and debuggerd.

Examples after:

  Fatal signal 13 (SIGPIPE), code 0 in tid 9157 (top)
  Fatal signal 11 (SIGSEGV), code 1, fault addr 0x0 in tid 9142 (crasher64)
  Fatal signal 6 (SIGABRT), code -6 in tid 9132 (crasher64)

(Note that the code still shows as 0 for SIGPIPE in the signal handler itself
but as -6 (SI_TKILL) in debuggerd; this is actually correct --- debuggerd is
showing the re-raised signal sent at the end of the signal handler that
initially showed the correct code 0.)

Change-Id: I71cad4ab61f422a4f6687a60ac770371790278e0

10 years agoFix for libgcc compat generation script.
Dmitriy Ivanov [Fri, 18 Apr 2014 23:03:03 +0000 (16:03 -0700)]
Fix for libgcc compat generation script.

  Taking into account possibility that external symbol
  could have been an OBJECT instead of function.

  b/14090368

Change-Id: Iac173d2dd1309ed53024306578137c26b1dbbf15