OSDN Git Service
Elliott Hughes [Fri, 18 Oct 2013 05:54:28 +0000 (05:54 +0000)]
Merge "Fix clone.S for x86_64."
Elliott Hughes [Fri, 18 Oct 2013 05:22:31 +0000 (22:22 -0700)]
Fix clone.S for x86_64.
We need to pull the arguments off the child stack and put them into the
registers they're expected to be in.
Change-Id: I064b3258cdd89d513c632857cabb46e52492af2b
Elliott Hughes [Fri, 18 Oct 2013 04:42:08 +0000 (04:42 +0000)]
Merge "Add missing aliases for off64_t functions in 64-bit land."
Elliott Hughes [Fri, 18 Oct 2013 04:33:29 +0000 (04:33 +0000)]
Merge "Add ffs and memcmp16 to x86_64."
Elliott Hughes [Fri, 18 Oct 2013 00:03:22 +0000 (17:03 -0700)]
Add ffs and memcmp16 to x86_64.
Change-Id: I652c1356f1c7c52299977181c2cf154386979380
Elliott Hughes [Thu, 17 Oct 2013 23:56:40 +0000 (16:56 -0700)]
Add missing aliases for off64_t functions in 64-bit land.
See the comment in SYSCALLS.TXT for an explanation.
Change-Id: I33d4056e84160c3cca74b7b588e9924a569753ed
Elliott Hughes [Thu, 17 Oct 2013 20:44:06 +0000 (20:44 +0000)]
Merge "Add SOCK_CLOEXEC and SOCK_NONBLOCK socket flags."
Andrei Emeltchenko [Mon, 14 Oct 2013 13:48:26 +0000 (16:48 +0300)]
Add SOCK_CLOEXEC and SOCK_NONBLOCK socket flags.
SOCK_CLOEXEC is used to atomically set close-on-exec flag for the new
descriptor(s), and SOCK_NONBLOCK is used to mark descriptor(s) as
non-blocking.
Change-Id: I8ba6a70543d23759e3ddcc7ff9c21b567184d681
Elliott Hughes [Thu, 17 Oct 2013 20:15:41 +0000 (20:15 +0000)]
Merge "Build the linker with the same -std flags as the rest of bionic."
Elliott Hughes [Thu, 17 Oct 2013 20:06:20 +0000 (13:06 -0700)]
Build the linker with the same -std flags as the rest of bionic.
Change-Id: I5bdb28ecfdf78a90f6b072077a1c5ffd63f5d2ae
Elliott Hughes [Thu, 17 Oct 2013 19:51:10 +0000 (19:51 +0000)]
Merge "x86_64: Add R_X86_64_64 relocation handling"
Elliott Hughes [Thu, 17 Oct 2013 19:45:35 +0000 (19:45 +0000)]
Merge "O_CLOEXEC is always defined for arm, mips, x86, and x86_64."
Elliott Hughes [Thu, 17 Oct 2013 18:45:22 +0000 (11:45 -0700)]
O_CLOEXEC is always defined for arm, mips, x86, and x86_64.
I built all four, just to be sure.
Change-Id: I5c818575c8142fccfc8fb163468832807a535e90
Elliott Hughes [Thu, 17 Oct 2013 18:38:31 +0000 (18:38 +0000)]
Merge "Fix sigaction(3) for 64-bit."
Elliott Hughes [Thu, 17 Oct 2013 05:27:54 +0000 (22:27 -0700)]
Fix sigaction(3) for 64-bit.
Also clean up <signal.h> and revert the hacks that were necessary
for 64-bit in linker/debugger.cpp until now.
Change-Id: I3b0554ca8a49ee1c97cda086ce2c1954ebc11892
Ben Cheng [Wed, 16 Oct 2013 23:52:17 +0000 (23:52 +0000)]
Merge "Refresh libc/kernel/common/linux/auxvec.h"
Ben Cheng [Wed, 16 Oct 2013 23:50:14 +0000 (16:50 -0700)]
Refresh libc/kernel/common/linux/auxvec.h
Change-Id: If9bb50a2f233cdb60e91bd979876edb917e0f66d
Ben Cheng [Wed, 16 Oct 2013 23:17:52 +0000 (23:17 +0000)]
Merge "Add processed uapi kernel headers (common and aarch64-specific)"
Ben Cheng [Wed, 16 Oct 2013 23:09:24 +0000 (16:09 -0700)]
Add processed uapi kernel headers (common and aarch64-specific)
Change-Id: If0be7b83bd8fe7cb02472d173f7c452aabf61124
Elliott Hughes [Wed, 16 Oct 2013 23:06:48 +0000 (23:06 +0000)]
Merge "Sort the syscalls.mk files, give all generated files the same header."
Ben Cheng [Wed, 16 Oct 2013 22:32:11 +0000 (22:32 +0000)]
Merge "Tweaked the cleanup scripts to handle uapi and aarch64 headers."
Ben Cheng [Wed, 16 Oct 2013 22:28:56 +0000 (15:28 -0700)]
Tweaked the cleanup scripts to handle uapi and aarch64 headers.
The processed uapi directory is now placed at libc/kernel/uapi as
opposed to libc/kernel/common/uapi as it contains
architectural-dependent headers now.
Change-Id: I53f814704a4d231b452fde398cd94257a0fb2eea
Elliott Hughes [Wed, 16 Oct 2013 21:27:59 +0000 (14:27 -0700)]
Sort the syscalls.mk files, give all generated files the same header.
No non-comment changes to the .S files.
Change-Id: Iafcfd004c3ea92b64268f80ab16df615b97cefac
Elliott Hughes [Wed, 16 Oct 2013 21:09:59 +0000 (21:09 +0000)]
Merge "Switch sigpending over to rt_sigpending."
Elliott Hughes [Wed, 16 Oct 2013 19:53:58 +0000 (12:53 -0700)]
Switch sigpending over to rt_sigpending.
Change-Id: I7b28984796b5fb343cfbcc47e0afc3a84293d417
Pavel Chupin [Wed, 16 Oct 2013 15:13:58 +0000 (19:13 +0400)]
x86_64: Add R_X86_64_64 relocation handling
Change-Id: I86ffc56fa6a9053bed44e92a579530c2beb8eb2c
Signed-off-by: Pavel Chupin <pavel.v.chupin@intel.com>
Elliott Hughes [Wed, 16 Oct 2013 02:51:34 +0000 (02:51 +0000)]
Merge "Remove support for the useless 'custom' option in SYSCALLS.TXT."
Elliott Hughes [Wed, 16 Oct 2013 01:18:58 +0000 (18:18 -0700)]
Remove support for the useless 'custom' option in SYSCALLS.TXT.
This cruft dates from a time when bionic would only output syscall
contants for the syscalls mentioned in SYSCALLS.TXT. I fixed that
a long time ago, but never followed through with the removal of what
was then confusingly called "stub" and was recently renamed "custom".
Change-Id: I8f3872a200b2dc8325e357cc5ee505ea4212ef95
Elliott Hughes [Wed, 16 Oct 2013 01:15:59 +0000 (01:15 +0000)]
Merge "Fix sigsuspend to use rt_sigsuspend on all platforms."
Elliott Hughes [Wed, 16 Oct 2013 01:01:56 +0000 (18:01 -0700)]
Fix sigsuspend to use rt_sigsuspend on all platforms.
Change-Id: I981c1a66d35480d4457a0a08a1b042dac94daa5b
Elliott Hughes [Wed, 16 Oct 2013 00:31:19 +0000 (00:31 +0000)]
Merge "'Avoid confusing "read prevented write" log messages' 2."
Elliott Hughes [Wed, 16 Oct 2013 00:17:05 +0000 (17:17 -0700)]
'Avoid confusing "read prevented write" log messages' 2.
This time it's assembler.
Change-Id: Iae6369833b8046b8eda70238bb4ed0cae64269ea
Elliott Hughes [Wed, 16 Oct 2013 00:09:54 +0000 (00:09 +0000)]
Merge "Make pthread_create report sched_setscheduler failures on LP64."
Elliott Hughes [Tue, 15 Oct 2013 23:53:45 +0000 (23:53 +0000)]
Merge "Avoid confusing "read prevented write" log messages."
Elliott Hughes [Tue, 15 Oct 2013 23:53:24 +0000 (23:53 +0000)]
Merge "Fix indentation in sched_getaffinity."
Elliott Hughes [Tue, 15 Oct 2013 23:51:17 +0000 (16:51 -0700)]
Make pthread_create report sched_setscheduler failures on LP64.
We couldn't fix this for 32-bit because there's too much broken
code out there. (Pretty much everyone asks for real-time
scheduling for all their threads, and the kernel says "don't be
stupid".)
Change-Id: I43c5271e6b6bb91278b9a19eec08cbf05391e3c4
Elliott Hughes [Tue, 15 Oct 2013 23:49:49 +0000 (16:49 -0700)]
Fix indentation in sched_getaffinity.
Change-Id: I966852149a0255ca132cd7a15f135a875ce400d0
Elliott Hughes [Tue, 15 Oct 2013 23:43:38 +0000 (16:43 -0700)]
Avoid confusing "read prevented write" log messages.
Moving to a "function: message" style avoids ambiguity.
Change-Id: If9d590e50265c61725d3673bd03796e65edd2d5e
Christopher Ferris [Tue, 15 Oct 2013 21:12:30 +0000 (21:12 +0000)]
Merge "Modify prefetch for krait memcpy."
Elliott Hughes [Tue, 15 Oct 2013 21:10:20 +0000 (21:10 +0000)]
Merge "Clean up the sigprocmask/pthread_sigmask implementation."
Christopher Ferris [Tue, 15 Oct 2013 19:10:06 +0000 (12:10 -0700)]
Modify prefetch for krait memcpy.
I originally modified the krait mainloop prefetch from cacheline * 8 to * 2.
This causes a perf degradation for copies bigger than will fit in the cache.
Fixing this back to the original * 8. I tried other multiples, but * 8 is th
sweet spot on krait.
Bug:
11221806
Change-Id: I1f75fad6440f7417e664795a6e7b5616f6a29c45
Elliott Hughes [Tue, 15 Oct 2013 18:23:57 +0000 (11:23 -0700)]
Clean up the sigprocmask/pthread_sigmask implementation.
Let's have both use rt_sigprocmask, like in glibc. The 64-bit ABIs
can share the same code as the 32-bit ABIs.
Also, let's test the return side of these calls, not just the
setting.
Bug:
11069919
Change-Id: I11da99f85b5b481870943c520d05ec929b15eddb
Stephen Hines [Fri, 11 Oct 2013 23:28:31 +0000 (23:28 +0000)]
Merge "Wrap sprintf()/snprintf() macros to prevent expansion errors."
Stephen Hines [Fri, 11 Oct 2013 23:27:40 +0000 (23:27 +0000)]
Merge "Fix clang warnings in bionic."
Stephen Hines [Fri, 11 Oct 2013 23:03:21 +0000 (16:03 -0700)]
Fix clang warnings in bionic.
This fixes a few diverse issues that clang warns on in bionic. First,
it specifies the appropriate converted types for format specifiers.
The "h" and "hh" modifiers specify that the user is passing a short or
char respectively. We were passing int deliberately in both cases and
relying on the compiler to implicitly downcast to the smaller type.
We also remove the non-standard "d" suffix from our double-precision
floating point constant. This is an extension for gcc that clang does
not implement. The third fix is to mark the c1 variable as unused,
since it truly is neither read nor written.
Change-Id: I4793352b9d3e58f1f4cac9e7581ef4b2a70b43c7
Stephen Hines [Fri, 11 Oct 2013 07:45:24 +0000 (00:45 -0700)]
Wrap sprintf()/snprintf() macros to prevent expansion errors.
Previously, FORTIFY_SOURCE used single macros to define these standard
functions for use with clang. This can cause conflicts with other macros used
to call these functions, particularly when those macros expand the number of
arguments to the function. This change wraps our macro definitions, so that
expansion properly takes place for programmer arguments first.
Change-Id: I55929b1fd2a643b9d14a17631c4bcab3b0b712cf
Nick Kralevich [Thu, 10 Oct 2013 04:11:48 +0000 (04:11 +0000)]
Merge "FORTIFY_SOURCE: fortify read()"
Elliott Hughes [Thu, 10 Oct 2013 04:08:43 +0000 (04:08 +0000)]
Merge "Clean up the cpuacct cruft."
Nick Kralevich [Thu, 10 Oct 2013 03:16:34 +0000 (20:16 -0700)]
FORTIFY_SOURCE: fortify read()
Change-Id: Ic7de163fe121db13e00560adb257331bc709814d
Elliott Hughes [Thu, 10 Oct 2013 00:35:36 +0000 (17:35 -0700)]
Clean up the cpuacct cruft.
Change-Id: I6ed63af8dfc2368e211420389fa8af4d5dc0908f
Elliott Hughes [Wed, 9 Oct 2013 23:51:28 +0000 (23:51 +0000)]
Merge "x86_64: Rename 64-bit linker to linker64"
Pavel Chupin [Wed, 6 Feb 2013 15:21:46 +0000 (19:21 +0400)]
x86_64: Rename 64-bit linker to linker64
That's for having both on the same system.
Change-Id: Ic2bc2c015e6486e8b6a7576f7b28d2d027534368
Signed-off-by: Pavel Chupin <pavel.v.chupin@intel.com>
Nick Kralevich [Wed, 9 Oct 2013 23:47:58 +0000 (23:47 +0000)]
Merge "Revert "FORTIFY_SOURCE: fortify read()""
Nick Kralevich [Wed, 9 Oct 2013 23:44:30 +0000 (16:44 -0700)]
Revert "FORTIFY_SOURCE: fortify read()"
This change reverts
*
fb3f956d075676c0438f2ee2bf3a5be659dfc04b.
*
65c99de2cb7a569ea17ca35e2f8f1e033421864b
Change-Id: Id5774eeede41130579115cf67a72ee914f2b47d5
Elliott Hughes [Wed, 9 Oct 2013 23:43:23 +0000 (23:43 +0000)]
Merge "Fix MIPS build."
Elliott Hughes [Wed, 9 Oct 2013 23:40:33 +0000 (16:40 -0700)]
Fix MIPS build.
Warnings are errors for all home-grown bionic code, and the arch-specific
code now counts as home-grown bionic code (it was mistakenly counted as
"not ours" before).
Change-Id: I9c6a881b0dc596bae7dfe112c5c189e073800a3a
Elliott Hughes [Wed, 9 Oct 2013 23:30:38 +0000 (23:30 +0000)]
Merge "Fix build."
Elliott Hughes [Wed, 9 Oct 2013 23:29:42 +0000 (16:29 -0700)]
Fix build.
'private' is no longer on the default include path inside bionic.
Change-Id: I9bfab213a496fac585787118603af3aa2a1f9951
Elliott Hughes [Wed, 9 Oct 2013 23:29:00 +0000 (23:29 +0000)]
Merge "Fix x86_64 build, clean up intermediate libraries."
Nick Kralevich [Wed, 9 Oct 2013 23:23:26 +0000 (23:23 +0000)]
Merge "FORTIFY_SOURCE: fortify read()"
Elliott Hughes [Wed, 9 Oct 2013 22:50:50 +0000 (15:50 -0700)]
Fix x86_64 build, clean up intermediate libraries.
The x86_64 build was failing because clone.S had a call to __thread_entry which
was being added to a different intermediate .a on the way to making libc.so,
and the linker couldn't guarantee statically that such a relocation would be
possible.
ld: error: out/target/product/generic_x86_64/obj/STATIC_LIBRARIES/libc_common_intermediates/libc_common.a(clone.o): requires dynamic R_X86_64_PC32 reloc against '__thread_entry' which may overflow at runtime; recompile with -fPIC
This patch addresses that by ensuring that the caller and callee end up in the
same intermediate .a. While I'm here, I've tried to clean up some of the mess
that led to this situation too. In particular, this removes libc/private/ from
the default include path (except for the DNS code), and splits out the DNS
code into its own library (since it's a weird special case of upstream NetBSD
code that's diverged so heavily it's unlikely ever to get back in sync).
There's more cleanup of the DNS situation possible, but this is definitely a
step in the right direction, and it's more than enough to get x86_64 building
cleanly.
Change-Id: I00425a7245b7a2573df16cc38798187d0729e7c4
Elliott Hughes [Wed, 9 Oct 2013 20:49:09 +0000 (20:49 +0000)]
Merge "Fix __errno for LP64 and clean up __get_tls."
Nick Kralevich [Wed, 9 Oct 2013 20:44:38 +0000 (13:44 -0700)]
FORTIFY_SOURCE: fortify read()
Change-Id: I3d7b4ec86d04efb865117ce7629a2e26917f3331
Elliott Hughes [Wed, 9 Oct 2013 01:50:24 +0000 (18:50 -0700)]
Fix __errno for LP64 and clean up __get_tls.
If __get_tls has the right type, a lot of confusing casting can disappear.
It was probably a mistake that __get_tls was exposed as a function for mips
and x86 (but not arm), so let's (a) ensure that the __get_tls function
always matches the macro, (b) that we have the function for arm too, and
(c) that we don't have the function for any 64-bit architecture.
Change-Id: Ie9cb989b66e2006524ad7733eb6e1a65055463be
Elliott Hughes [Wed, 9 Oct 2013 01:41:45 +0000 (01:41 +0000)]
Merge "Fix 32-bit build."
Elliott Hughes [Wed, 9 Oct 2013 01:30:44 +0000 (18:30 -0700)]
Fix 32-bit build.
libc/tzcode/localtime.c: In function 'differ_by_repeat':
libc/tzcode/localtime.c:338:2: error: comparison is always false due to limited range of data type [-Werror=type-limits]
Change-Id: Ic84be6391a66e9d50ed98f41d865387c77a60ffa
Elliott Hughes [Wed, 9 Oct 2013 00:20:08 +0000 (00:20 +0000)]
Merge "Fix malloc debugging for LP64."
Elliott Hughes [Wed, 9 Oct 2013 00:18:50 +0000 (00:18 +0000)]
Merge "Don't allow int<->pointer conversions."
Elliott Hughes [Wed, 9 Oct 2013 00:18:44 +0000 (00:18 +0000)]
Merge "Fix bionic's built-in stack trace dumping for LP64."
Elliott Hughes [Wed, 9 Oct 2013 00:04:33 +0000 (17:04 -0700)]
Don't allow int<->pointer conversions.
Normally we don't have -Werror for upstream code, but for those warnings
that probably point to 32-bit assumptions about pointers, we want those
warnings to always be errors.
Change-Id: Ibece9caf09b2f7989ca600ef448d07868669a8fb
Elliott Hughes [Wed, 9 Oct 2013 00:02:26 +0000 (17:02 -0700)]
Fix bionic's built-in stack trace dumping for LP64.
Change-Id: I967c5789d7bb2d3d248d94d81a40d5ec4e1bf26d
Elliott Hughes [Tue, 8 Oct 2013 23:16:01 +0000 (16:16 -0700)]
Fix malloc debugging for LP64.
Change-Id: Idd0b239f5c66d45de315d556271a5d13b8eb907c
Elliott Hughes [Tue, 8 Oct 2013 23:09:54 +0000 (23:09 +0000)]
Merge "Fix the ALIGN and ALIGNBYTES macros for LP64."
Elliott Hughes [Tue, 8 Oct 2013 22:31:44 +0000 (15:31 -0700)]
Fix the ALIGN and ALIGNBYTES macros for LP64.
Change-Id: I3a361255afce375ab1cefa449721f0aea4d47919
Elliott Hughes [Tue, 8 Oct 2013 22:02:20 +0000 (22:02 +0000)]
Merge "pthread_exit should call __NR_exit with status 0."
Elliott Hughes [Tue, 8 Oct 2013 21:50:34 +0000 (21:50 +0000)]
Merge "Use /system/lib64 and /vendor/lib64 for 64-bit libraries."
Elliott Hughes [Tue, 8 Oct 2013 21:48:05 +0000 (14:48 -0700)]
pthread_exit should call __NR_exit with status 0.
We shouldn't have been passing the bottom 32 bits of the address used
for pthread_join to the kernel.
Change-Id: I487e5002d60c27adba51173719213abbee0f183f
Elliott Hughes [Tue, 8 Oct 2013 21:27:10 +0000 (14:27 -0700)]
Use /system/lib64 and /vendor/lib64 for 64-bit libraries.
Change-Id: I4886aeb3070bf97b4cfe8053388ecb1bda288017
Elliott Hughes [Tue, 8 Oct 2013 20:42:35 +0000 (20:42 +0000)]
Merge "Make logging fall back to /dev/stderr if we're on the host."
Elliott Hughes [Tue, 8 Oct 2013 20:19:00 +0000 (13:19 -0700)]
Make logging fall back to /dev/stderr if we're on the host.
Otherwise you get no logging, which sucks.
Change-Id: Iea1e8f996461afbb217a55711b7967005c39cfcb
Elliott Hughes [Tue, 8 Oct 2013 18:37:22 +0000 (18:37 +0000)]
Merge "Use linker64 for the 64-bit linker."
Elliott Hughes [Tue, 8 Oct 2013 17:05:05 +0000 (10:05 -0700)]
Use linker64 for the 64-bit linker.
Change-Id: I13fc7f93274f99e4cf99b077afdf5293e7233f39
Elliott Hughes [Tue, 8 Oct 2013 17:01:41 +0000 (17:01 +0000)]
Merge "Use C99 structure initializer designator style."
Elliott Hughes [Tue, 8 Oct 2013 16:39:52 +0000 (09:39 -0700)]
Use C99 structure initializer designator style.
clang warns about using the GCC style of designator.
Change-Id: I86ec79f06c8774618082859f48d7d1f576520e32
Elliott Hughes [Tue, 8 Oct 2013 16:59:16 +0000 (16:59 +0000)]
Merge "x86_64 linker."
Elliott Hughes [Sat, 5 Oct 2013 00:01:33 +0000 (17:01 -0700)]
x86_64 linker.
Based on I8dc3e2cb596f75dc58ae82e4dc58f8c177dd3323 by
Pavel Chupin <pavel.v.chupin@intel.com>.
Change-Id: Icd582d277cbe273477b450f2848343d72c86ec9f
Elliott Hughes [Tue, 8 Oct 2013 16:53:02 +0000 (16:53 +0000)]
Merge "Refactor the syscall generation script."
Elliott Hughes [Tue, 8 Oct 2013 16:12:53 +0000 (16:12 +0000)]
Merge "Fix Clang warning since KernelArgumentBlock is actually a class."
Stephen Hines [Tue, 8 Oct 2013 07:05:44 +0000 (00:05 -0700)]
Fix Clang warning since KernelArgumentBlock is actually a class.
Change-Id: Id72868d80feffbbc5f7d1e43beaed5d4e4d95f52
Elliott Hughes [Tue, 8 Oct 2013 06:53:13 +0000 (23:53 -0700)]
Refactor the syscall generation script.
Primarily so that the new x86_64 alias functionality is now available for
all architectures.
Change-Id: I9fde59093a1d08de98923f121a6e3d05ec5801d2
Elliott Hughes [Tue, 8 Oct 2013 06:03:01 +0000 (06:03 +0000)]
Merge "Add an optional alias list to SYSCALLS.TXT"
Elliott Hughes [Mon, 7 Oct 2013 18:33:51 +0000 (18:33 +0000)]
Merge "libc: fix __cxa_atexit implicit declaration"
synergydev [Mon, 7 Oct 2013 08:13:11 +0000 (01:13 -0700)]
libc: fix __cxa_atexit implicit declaration
Change-Id: Iacad18b332a717e4485c83df4bd42a850ff7699f
Elliott Hughes [Mon, 7 Oct 2013 17:49:18 +0000 (17:49 +0000)]
Merge "Clean up the x86 and x86_64 _exit_with_stack_teardown implementations."
H.J. Lu [Fri, 4 Oct 2013 17:03:17 +0000 (10:03 -0700)]
Add an optional alias list to SYSCALLS.TXT
This patch adds an optional alias list to SYSCALLS.TXT. It is used to
create aliases for a syscall. For x86-64, lseek64 is an alias for lseek.
Change-Id: Icb11fd2bb461ea4f5f0a26bfc585471d7d7cc468
Signed-off-by: H.J. Lu <hjl.tools@gmail.com>
Signed-off-by: Pavel Chupin <pavel.v.chupin@intel.com>
Elliott Hughes [Mon, 7 Oct 2013 17:27:17 +0000 (17:27 +0000)]
Merge "Don't define the meaningless SOFTFLOAT for x86."
Elliott Hughes [Mon, 7 Oct 2013 17:20:24 +0000 (10:20 -0700)]
Clean up the x86 and x86_64 _exit_with_stack_teardown implementations.
Change-Id: I4bcbbc53893612bd94643ef07722becb00f91792
Elliott Hughes [Mon, 7 Oct 2013 16:53:44 +0000 (09:53 -0700)]
Don't define the meaningless SOFTFLOAT for x86.
Change-Id: I9f932ad5d9f731a0de18efb881d02cedf9fcbede
Elliott Hughes [Mon, 7 Oct 2013 05:32:56 +0000 (05:32 +0000)]
Merge "Fix MIPS build."
Elliott Hughes [Mon, 7 Oct 2013 05:29:07 +0000 (22:29 -0700)]
Fix MIPS build.
Although 'register' is deprecated, we need to use v1, and there's
no way to do that through register constraints on the assembler
fragment itself.
Change-Id: Ib5b12c4c3652513d10cc61d4a4b11314ece25663