Let the assembler handle the low-level details of generating .eh_frame,
and use .cfi_escape to output the DWARF3 DW_CFA_def_cfa_expression and
DW_CFA_expression instructions.
Explicitly output a nop instruction between the FDE start
(.cfi_startproc) and the trampoline symbol.
x86_64: remove the rsp instruction, which is redundant with the CFA
itself.
Bug: http://b/169383888
Test: compare `objdump -Wf` output before and after
Test: bionic-unit-tests
Change-Id: Ie2a6d111cb473596be8c9f4fd64534e91d88f2a1
Our various fd debugging facilities get extremely confused by a vforked
process closing file descriptors in preparation to exec: fdsan can
abort, and fdtrack will delete backtraces for any file descriptors that
get closed. Keep track of whether we're in a vforked child in order to
be able to detect this.
Bug: http://b/153926671
Test: 32/64-bit bionic-unit-tests on blueline, x86_64 emulator
Change-Id: I8a082fd06bfdfef0e2a88dbce350b6f667f7df9f
Switch the __x86.get_pc_thunk.{bx,cx} functions from .gnu.linkonce to
comdat section groups. lld doesn't implement .gnu.linkonce and will
instead discard the input sections. Currently, it might produce a
faulty binary that has no get_pc_thunk function in it, which would
crash.
Normally, these functions are generated by GCC and are present in many
object files compiled with PIC. Clang doesn't use them, and instead
initializes the PIC base register with a "call 1f; 1: pop %ebx" pair.
Bug: http://b/154376560
Bug: https://bugs.llvm.org/show_bug.cgi?id=45594
Test: bionic-unit-tests
Change-Id: I3e4857aecdc281b32232270bff4880433e5b6afa
There are places in frameworks and art code that directly included
private bionic header files. Move these files to the new platform
include files.
This change also moves the __get_tls.h header file to tls.h and includes
the tls defines header so that there is a single header that platform
code can use to get __get_tls and the defines.
Also, simplify the visibility rules for platform includes.
Bug: 141560639
Test: Builds and bionic unit tests pass.
Change-Id: I9e5e9c33fe8a85260f69823468bc9d340ab7a1f9
Merged-In: I9e5e9c33fe8a85260f69823468bc9d340ab7a1f9
(cherry picked from commit 44631c919a)
With the switch to libcrt.builtins, some symbols no longer becomes
exported. Add dummy references to them to force them to be exported.
This is to maintain backwards binary compatibility with ancient Android
versions.
x86 and x86_64 crashes with libcrt, keep using libgcc for now.
Test: bionic-unit-tests
Bug: 29275768
Bug: 122993569
Change-Id: Ieab5af354e3924af4a03d888b28c6e75090cb613
When the linker is invoked on itself, (`linker64 /system/bin/linker64`),
the linker prints an error, because self-invocation isn't allowed. The
current method for detecting self-invocation fails because the second
linker instance can crash in a constructor function before reaching
__linker_init.
Fix the problem by moving the error check into a constructor function,
which finishes initializing libc sufficiently to call async_safe_fatal.
The only important thing missing is __libc_sysinfo on 32-bit x86. The aux
vector isn't readily accessible, so use the fallback int 0x80.
Bug: http://b/123637025
Test: bionic unit tests (32-bit x86)
Change-Id: I8be6369e8be3938906628ae1f82be13e6c510119
Instead of passing the address of a KernelArgumentBlock to libc.so for
initialization, use __loader_shared_globals() to initialize globals.
Most of the work happened in the previous CLs. This CL switches a few
KernelArgumentBlock::getauxval calls to [__bionic_]getauxval and stops
routing the KernelArgumentBlock address through the libc init functions.
Bug: none
Test: bionic unit tests
Change-Id: I96c7b02c21d55c454558b7a5a9243c682782f2dd
Merged-In: I96c7b02c21d55c454558b7a5a9243c682782f2dd
(cherry picked from commit 746ad15912)
__libc_sysinfo is hidden, so accessing it doesn't require a relocated GOT.
It is important not to have a relocatable initializer on __libc_sysinfo,
because if it did have one, and if we initialized it before relocating the
linker, then on 32-bit x86 (which uses REL rather than RELA), the
relocation step would calculate the wrong addend and overwrite
__libc_sysinfo with garbage.
Asides:
* It'd be simpler to keep the __libc_sysinfo initializer for static
executables, but the loader pulls in libc_init_static (even though it
uses almost none of the code in that file, like __libc_init).
* The loader has called __libc_init_sysinfo three times by the time it
has relocated itself. A static executable calls it twice, while libc.so
calls it only once.
Bug: none
Test: lunch aosp_x86-userdebug ; emulator
Test: adb shell /data/nativetest/bionic-unit-tests/bionic-unit-tests
Test: adb shell /data/nativetest/bionic-unit-tests-static/bionic-unit-tests-static
Change-Id: I5944f57847db7191608f4f83dde22b49e279e6cb
Our arm and x86 ABIs shipped with a 32-bit `sigset_t`, so we need to
use sigprocmask64 to save/restore the RT signals too. (This is important
because several are in use by the system, and the behavior of our 32-bit
`sigset_t` is to clear the RT signals.)
Bug: http://b/72493232
Test: ran tests
Change-Id: Idff91f8b2849276e5a3073d07eccd84e02a34410
No effect right now, because sigprocmask on LP32 also only touches
the non-RT signals, but this makes it easier to switch to __rt_sigprocmask.
Bug: http://b/72460436
Test: ran tests
Change-Id: I693f0ea36701e9ab5d10e6aefb26387ba45a6064
These sources are going to be used for the NDK as well, and the NDK
still uses GCC.
This partially reverts commit
4af220cfef. That commit also removed
this symbol from the linker's crtbegin (it has its own). That is
still only built with Clang, so we don't need to revert that part.
Test: treehugger
Bug: None
Change-Id: Iba231baf298e228135bdf48dfed87f9089975eb1
To make it easier for Native Bridge implementations
to override these symbols.
Bug: http://b/67993967
Test: make
Change-Id: I4c53e53af494bca365dd2b3305ab0ccc2b23ba44
The kernel does seem to read some of the "unused" fields if only to do
basic validity checks.
valgrind complained about this, and I'm not convinced it's spurious.
Bug: http://b/38034461
Test: ran tests
Change-Id: I077fcc75bf0738491242166fc4576a9693ce18a5
This patch uses __kernel_vsyscall instead of "int 0x80"
as the syscall entry point. AT_SYSINFO points to
an adapter to mask the arch specific difference and gives a
performance boost on i386 architecture.
Change-ID: Ib340c604d02c6c25714a95793737e3cfdc3fc5d7
Signed-off-by: Mingwei Shi <mingwei.shi@intel.com>
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>
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
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
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
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
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
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
