This change provides __restore/__restore_rt on x86 and __restore_rt on
x86_64 with unwinding information to be able to unwind through signal
frame via libgcc provided unwinding interface. See comments inlined for
more details.
Also remove the test that had a dependency on
__attribute__((cleanup(foo_cleanup))). It doesn't provide us with any
better test coverage than we have from the newer tests, and it doesn't
work well across a variety architectures (presumably because no one uses
this attribute in the real world).
Tested this on host via bionic-unit-tests-run-on-host on both x86 and
x86-64.
Bug: 17436734
Change-Id: I2f06814e82c8faa732cb4f5648868dc0fd2e5fe4
Signed-off-by: Pavel Chupin <pavel.v.chupin@intel.com>
Add the missing prototypes, fix the existing prototypes to use clockid_t
rather than int, fix clock_nanosleep's failure behavior, and add simple
tests.
Bug: 17644443
Bug: https://code.google.com/p/android/issues/detail?id=77372
Change-Id: I03fba369939403918abcabae9551a7123953d780
Signed-off-by: Haruki Hasegawa <h6a.h4i.0@gmail.com>
For silvermont, the __popcountsi2 symbol does not get exported by libc.
But for atom, this symbol is exported. Since we already exported this symbol
for previous releases, it's better to just follow through and force
the export, but only for 32 bit. x86 64 bit will not export this symbol.
Bug: 17681440
(cherry picked from commit d11eac3455)
Change-Id: I93704c721d98d569922f606f214069bda24872ba
* LP32 should use sa_restorer too. gdb expects this, and future (>= 3.15) x86
kernels will apparently stop supporting the case where SA_RESTORER isn't
set.
* gdb and libunwind care about the exact instruction sequences, so we need to
modify the code slightly in a few cases to match what they're looking for.
* gdb also cares about the exact function names (for some architectures),
so we need to use __restore and __restore_rt rather than __sigreturn and
__rt_sigreturn.
* It's possible that we don't have a VDSO; dl_iterate_phdr shouldn't assume
that getauxval(AT_SYSINFO_EHDR) will return a non-null pointer.
This fixes unwinding through a signal handler in gdb for all architectures.
It doesn't fix libunwind for arm and arm64. I'll keep investigating that...
Bug: 17436734
Change-Id: Ic1ea1184db6655c5d96180dc07bcc09628e647cb
There are number of changes in the way IFUNC related relocations are done:
1. IRELATIVE relocations are now supported for x86/x86_64 and arm64.
2. IFUNC relocations are now relying on static linker to generate
them in correct order - this removes necessety of additional
relocation pass for ifuncs.
3. Related to 2: rela?.dyn relocations are preformed before .plt ones.
4. Ifunc are resolved on symbol lookup this approach allowed to avoid
mprotect(PROT_WRITE) call on r-x program segments.
Bug: 17399706
Bug: 17177284
Change-Id: I414dd3e82bd47cc03442c5dfc7c279949aec51ed
The use of the .hidden directive to avoid going via the PLT for
__set_errno had the side-effect of actually making __set_errno
hidden (which is odd because assembler directives don't usually
affect symbols defined in a different file --- you can't even
create a weak reference to a symbol that's defined in a different
file).
This change switches the system call stubs over to a new always-hidden
__set_errno_internal and has a visible __set_errno on LP32 just for
binary compatibility with old NDK apps.
(cherry-pick of 7efad83d430f4d824f2aaa75edea5106f6ff8aae.)
Bug: 17423135
Change-Id: I6b6d7a05dda85f923d22e5ffd169a91e23499b7b
On most architectures the kernel subtracts a random offset to the stack
pointer in create_elf_tables by calling arch_align_stack before writing
the auxval table and so on. On all but x86 this doesn't cause a problem
because the random offset is less than a page, but on x86 it's up to two
pages. This means that our old technique of rounding the stack pointer
doesn't work. (Our old implementation of that technique was wrong too.)
It's also incorrect to assume that the main thread's stack base and size
are constant. Likewise to assume that the main thread has a guard page.
The main thread is not like other threads.
This patch switches to reading /proc/self/maps (and checking RLIMIT_STACK)
whenever we're asked.
Bug: 17111575
Signed-off-by: Fengwei Yin <fengwei.yin@intel.com>
Change-Id: I1d4dbffe7bc7bda1d353c3a295dbf68d29f63158
The old definition only worked for functions that didn't use numbered
local labels. Upstream uses '666' not only as some kind of BSD in-joke,
but also because there's little likelihood of any function having
labels that high.
There's a wider question about whether we actually want to go via the
PLT at all in this code, but that's a question for another day.
(cherry-pick of 72d7e667c7e926cb120c4edb53cbf74c652ab915.)
Bug: 16906712
Change-Id: I3cd8ecc448b33f942bb6e783931808ef39091489
Clean up the x86/x86_64 assembler. The motivator (other than reducing
confusion) was that asm.h incorrectly checked PIC rather than __PIC__.
Bug: 16823325
Change-Id: Iaa9d45009e93a4b31b719021c93ac221e336479b
Also fix a few formatting issues in copyright headers that were confusing
the script (though obviously it would be better if the script were smarter).
Change-Id: I7f561bef4f84fdcbd84f375ee226bd65db0e507b
In practice, with this implementation we never need to make a system call.
We get the main thread's tid (which is the same as our pid) back from
the set_tid_address system call we have to make during initialization.
A new pthread will have the same pid as its parent, and a fork child's
main (and only) thread will have a pid equal to its tid, which we get for
free from the kernel before clone returns.
The only time we'd actually have to make a getpid system call now is if
we take a signal during fork and the signal handler calls getpid. (That,
or we call getpid in the dynamic linker while it's still dealing with its
own relocations and hasn't even set up the main thread yet.)
Bug: 15387103
Change-Id: I6d4718ed0a5c912fc75b5f738c49a023dbed5189
The C library didn't export the 'index' symbol, but its C++ name-mangling
instead, which broke the ABI and prevented some applications from loading
properly.
The main reason was that the implementation under bionic/index.cpp relied
on the declaration to specify that the function has C linkage.
However, the declaration for index() was removed from both <string.h>
and <strings.h> in a recent patch, which made the compiler think it was
ok to compile the function with C++ linkage instead!
This patch does the following:
- Move index() definition to bionic/ndk_cruft.cpp and ensure it uses
C linkage.
Note that this removes index() from the 64-bit library entirely, this
is intentional and will break source compatibility. Simply replacing
an index() call with the equivalent strchr() should be enough to fix
this in third-party code.
- Remove bionic/index.cpp from the tree and build files.
- Remove x86 assembly implementation from arch-x86/ to avoid conflict
with the one in ndk_cruft.cpp
BUG=15606653
Change-Id: I816b589f69c8f8a6511f6be6195d20cf1c4e8123
These were both removed from POSIX 2004, and we don't define an
implementation for getw(3). Keep the definition of put(3) on LP32 for
binary compatibility.
Bug: 13935372
Change-Id: Iba384b45093ac6d2d7c2d81f7980cd7701dd6f56
vfork() was removed from POSIX 2008, so this replaces its implementation
with a call to fork().
Bug: 13935372
Change-Id: I6d99ac9e52a2efc5ee9bda1cab908774b830cedc
__set_errno returns -1 exactly so that callers don't need to bother.
The other architectures were already taking advantage of this, but
no one had ever fixed x86 and x86_64.
Change-Id: Ie131494be664f6c4a1bbf8c61bbbed58eac56122
x86-64 needs these CFI directives to stop unwinding here.
I've also cleaned up the assembler a little, and made x86 and x86-64
a little more alike.
Bug: 15195760
Change-Id: I40f92c007843c29c933bb6876fe2b4611e1b946b
Introduce a test for memmove that catches a fault.
Fix both 32- and 64-bit versions of slm-tuned memmove.
Change-Id: Ib416def2610a0972e32c3b9b6055b54967643dc3
Signed-off-by: Varvara Rainchik <varvara.rainchik@intel.com>
System calls can be pretty slow. This is mako, which has one of our
lowest latencies:
iterations ns/op
BM_unistd_getpid 10000000 209
BM_unistd_gettid 200000000 8
Bug: 15297299 (kernel panic from too many gettid calls)
Bug: 15315766 (excessive gettid overhead in liblogd)
Change-Id: I49656c0fc5b5d092390264a59e4f2c0d8a8b1aeb
The syscall generation always used 4 bytes for each push cfi directive.
However, the first push should always use an offset of 8 bytes, each
subsequent push after that is only 4 bytes though.
Change-Id: Ibaabd107f399ef67010b9a08213783957c2f74a9
The problem with the original patch was that using syscall(3) means that
errno can be set, but pthread_create(3) was abusing the TLS errno slot as
a pthread_mutex_t for the thread startup handshake.
There was also a mistake in the check for syscall failures --- it should
have checked against -1 instead of 0 (not just because that's the default
idiom, but also here because futex(2) can legitimately return values > 0).
This patch stops abusing the TLS errno slot and adds a pthread_mutex_t to
pthread_internal_t instead. (Note that for LP64 sizeof(pthread_mutex_t) >
sizeof(uintptr_t), so we could potentially clobber other TLS slots too.)
I've also rewritten the LP32 compatibility stubs to directly reuse the
code from the .h file.
This reverts commit 75c55ff84e.
Bug: 15195455
Change-Id: I6ffb13e5cf6a35d8f59f692d94192aae9ab4593d
This reverts commit ced906c849.
Causes issues on art / dalvik due to a broken return value
check and other undiagnosed issues.
bug: 15195455
Change-Id: I5d6bbb389ecefb0e33a5237421a9d56d32a9317c
Add optimized versions of bcopy and wmemmove for AArch64 based on the
memmove implementation
Change-Id: I82fbe8a7221ce224c567ffcfed7a94a53640fca8
Signed-off-by: Bernhard Rosenkraenzer <Bernhard.Rosenkranzer@linaro.org>
This reverts commit 8167dd7cb9.
For some reason I thought the bcopy change was bzero. The bcopy code doesn't pass our tests, so reverting until I can figure out what's wrong.
Change-Id: Id89fe959ea5105cd58dff6bba8d91a30cc4bcb07
Add optimized versions of bcopy and wmemmove for AArch64 based on the
memmove implementation
Change-Id: Ie43d0ff4f8ec4edba5b4fb5ccacd941f81ac6557
Signed-off-by: Bernhard Rosenkraenzer <Bernhard.Rosenkranzer@linaro.org>
glibc doesn't have tkill or tgkill and says "use syscall(3) instead".
I've left tgkill since it's quite widely used, but there's no reason
to have tkill as well.
Bug: 11156955
Change-Id: Ifc0af750320086f829bc9914551c172b501f3b60
Also hide part of the system properties compatibility code, since
we needed to touch that to keep it building.
I'll remove __futex_syscall4 and futex in a later patch.
Bug: 11156955
Change-Id: Ibbf42414c5bb07fb9f1c4a169922844778e4eeae
This was accidentally added at a time when you couldn't add a constant
to <syscall.h> without generating an assembly stub! (You no longer need
to add the constants at all.)
Bug: 11156955
Change-Id: I053c17879138787976c744a5ecf7d30ee51dc48f
The library exists outside bionic. It is dynamically loaded, to replace selected
standard socket syscalls with versions that talk to netd.
Change connect() to use the library if available.
(cherry picked from commit 3a6b627a14df8111b03e452f2df4b5f4938e0e49)
Change-Id: Ib6198e19dbc306521a26fcecfdf6e8424d163fc9
Add following functions:
bcopy, memcpy, memmove, memset, bzero, memcmp, wmemcmp, strlen,
strcpy, strncpy, stpcpy, stpncpy.
Create new directories inside arch-x86 to specify architecture: atom,
silvermont and generic (non atom or silvermont architectures are treated like generic).
Due to introducing optimized versions of stpcpy and stpncpy,
c-implementations of these functions are moved from
common for architectures makefile to arm and mips specific makefiles.
Change-Id: I990f8061c3e9bca1f154119303da9e781c5d086e
Signed-off-by: Varvara Rainchik <varvara.rainchik@intel.com>
Also let clone(2) set the TLS for x86.
Also ensure we initialize the TLS before we clone(2) for all architectures.
Change-Id: Ie5fa4466e1c9ee116a281dfedef574c5ba60c0b5
Also ensure that arm64/x86-64/x86 assembler uses local labels.
(There are are so many non-local labels in arm that fixing them
seems out of scope.)
Also synchronize the __bionic_clone.S comments.
Change-Id: I03b4f84780d996b54d6637a074638196bbb01cd4
clone(2) is the public symbol.
Also switch a test from __bionic_clone to clone; testing public API
means the test now works on glibc too.
Change-Id: If59def26a00c3afadb8a6cf9442094c35a59ffde
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
The upstream intention was for this to be architecture-dependent, but it's a
lot clearer if we just have one copy.
Change-Id: I4e8310496145f9f411cd2e847c8cd023b1d758e9
Also move isinf and isnan into libc like everyone else.
Also move fpclassify to libc like the BSDs (but unlike glibc). We need
this to be able to upgrade our float/double/long double parsing to gdtoa.
Also add some missing aliases. We now have all of:
isnan, __isnan, isnanf, __isnanf, isnanl, __isnanl,
isinf, __isinf, isinff, __isinff, isinfl, __isinfl,
__fpclassify, __fpclassifyd, __fpclassifyf, __fpclassifyl.
Bug: 13469877
Change-Id: I407ffbac06c765a6c5fffda8106c37d7db04f27d
Adds Silvermont specific cache sizes for bionic optimizations.
Change-Id: Ib992f530b8c485121b2874470fd6bed2212adb0f
Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
the POSIX standard is that SIZE_MAX is defined
in stdint.h, not limits.h.
Change-Id: Iafd8ec71d1840541feaca4f53b2926b398293fac
Signed-off-by: Webb, Russell <russell.webb@intel.com>
Signed-off-by: Fengwei Yin <fengwei.yin@intel.com>
Reviewed-by: Ross, Andrew J <andrew.j.ross@intel.com>
Reviewed-by: Boie, Andrew P <andrew.p.boie@intel.com>
Reviewed-by: Gumbel, Matthew K <matthew.k.gumbel@intel.com>
Tested-by: Gumbel, Matthew K <matthew.k.gumbel@intel.com>
Also add the corresponding constant, struct, and function declarations
to <sys/socket.h>, and perfunctory tests so we know that the symbols
actually exist.
Signed-off-by: Guillaume Ranquet <guillaumex.ranquet@intel.com>
Change-Id: Ib0d854239d3716be90ad70973c579aff4895a4f7
Our <machine/asm.h> files were modified from upstream, to the extent
that no architecture was actually using the upstream ENTRY or END macros,
assuming that architecture even had such a macro upstream. This patch moves
everyone to the same macros, with just a few tweaks remaining in the
<machine/asm.h> files, which no one should now use directly.
I've removed most of the unused cruft from the <machine/asm.h> files, though
there's still rather a lot in the mips/mips64 ones.
Bug: 12229603
Change-Id: I2fff287dc571ac1087abe9070362fb9420d85d6d
This gives us:
* <dirent.h>
struct dirent64
readdir64, readdir64_r, alphasort64, scandir64
* <fcntl.h>
creat64, openat64, open64.
* <sys/stat.h>
struct stat64
fstat64, fstatat64, lstat64, stat64.
* <sys/statvfs.h>
struct statvfs64
statvfs64, fstatvfs64.
* <sys/vfs.h>
struct statfs64
statfs64, fstatfs64.
This also removes some of the incorrect #define hacks we've had in the
past (for stat64, for example, which we promised to clean up way back
in bug 8472078).
Bug: 11865851
Bug: 8472078
Change-Id: Ia46443521918519f2dfa64d4621027dfd13ac566
1. Moved arch-specific setup to their own files:
- <arch>/<arch>.mk, arch-specific configs. Variables in those config
end with the arch name.
- removed the extra complexity introduced by function libc-add-cpu-variant-src,
which seems to be not very useful these days.
2. Separated out the crt object files generation rules and set up the
rules for both TARGET_ARCH and TARGET_2ND_ARCH.
3. Build all the libraries for both TARGET_ARCH and TARGET_2ND_ARCH,
with the arch-specific LOCAL_ variables.
Bug: 11654773
Change-Id: I9c2d85db0affa49199d182236d2210060a321421
Remove the linker's reliance on BSD cruft and use the glibc-style
ElfW macro. (Other code too, but the linker contains the majority
of the code that needs to work for Elf32 and Elf64.)
All platforms need dl_iterate_phdr_static, so it doesn't make sense
to have that part of the per-architecture configuration.
Bug: 12476126
Change-Id: I1d7f918f1303a392794a6cd8b3512ff56bd6e487
Most of <machine/_types.h> was either unused, wrong, or identical across
all 32-/64-bit architectures.
I'm not a huge fan of <sys/_types.h> either, but moving the bits we need
up into there is a step forward.
Bug: 12213562
Change-Id: Id13551c78966e324beee2dd90c5575e37d2a71e6
The situation here is a bit confusing. On 64-bit, rlimit and rlimit64 are
the same, and so getrlimit/getrlimit64, setrlimit/setrlimit64,
and prlimit/prlimit64 are all the same. On 32-bit, rlimit and rlimit64 are
different. 32-bit architectures other than MIPS go one step further by having
an even more limited getrlimit system call, so arm and x86 need to use
ugetrlimit instead of getrlimit. Worse, the 32-bit architectures don't have
64-bit getrlimit- and setrlimit-equivalent system calls, and you have to use
prlimit64 instead. There's no 32-bit prlimit system call, so there's no
easy implementation of that --- what should we do if the result of prlimit64
won't fit in a struct rlimit? Since 32-bit survived without prlimit/prlimit64
for this long, I'm not going to bother implementing prlimit for 32-bit.
We need the rlimit64 functions to be able to build strace 4.8 out of the box.
Change-Id: I1903d913b23016a2fc3b9f452885ac730d71e001
Modify the syscalls script to generate the cfi directives for x86
syscalls.
Update the x86 syscalls.
Change-Id: Ia1993dc714a7e79f917087fff8200e9a02c52603
This patch switches to using the uapi constants. It also adds the missing
setns system call, fixes sched_getcpu's error behavior, and fixes the
gensyscalls script now ARM is uapi-only too.
Change-Id: I8e16b1693d6d32cd9b8499e46b5d8b0a50bc4f1d
Also make the other architectures more similar to one another,
use NULL instead of 0 in calling code, and remove an unused #define.
Change-Id: I52b874afb6a351c802f201a0625e484df6d093bb
