Also mark the non-mangled fcontext asm function symbols
as hidden visibility
These functions should not be exported as dynamic
symbols by boost, only the namespaced C++ symbols
introduced by this patch should.
This implements the required asm functions for sparc64_sysv_elf and
was tested on an OpenBSD/sparc64 system.
Jump_fcontext uses an extra C call frame to store the frame-pointer and
return address. Because of this the code is simply a save (to new reg
window), then forcing a window flush and finally switch stack and restore
from there.
Since jump_fcontext() uses a register window and stack frame, make_fcontext()
reserves two call frames on the stack (one for jump_fcontext() and the other
for the callback function).
OpenBSD/sparc64 uses stackghost which prevents userland from overriding the
return-address on the stack. Because of this make_fcontext() uses an extra
trampoline to implement the _exit(0) call if the callback returns.
All tests pass with this on OpenBSD/sparc64 (also the tests for fiber,
coroutine and coroutine2).
This implements the required asm functions for sparc64_sysv_elf and
was tested on an OpenBSD/sparc64 system.
Jump_fcontext uses an extra C call frame to store the frame-pointer and
return address. Because of this the code is simply a save (to new reg
window), then forcing a window flush and finally switch stack and restore
from there.
Since jump_fcontext() uses a register window and stack frame, make_fcontext()
reserves two call frames on the stack (one for jump_fcontext() and the other
for the callback function).
OpenBSD/sparc64 uses stackghost which prevents userland from overriding the
return-address on the stack. Because of this make_fcontext() uses an extra
trampoline to implement the _exit(0) call if the callback returns.
All tests pass with this on OpenBSD/sparc64 (also the tests for fiber,
coroutine and coroutine2).
`in` operator in bjam always returns true if its first
argument has no elements[1]. This means that if `os.platform`
is empty (not detected), the construction introduced
in commit d039c8e4da sets
ABI to `aapcs` on all platforms where `os.platform` is
empty, including, e.g. riscv64, and breaks build there.
This commit refactors the condition to use '=' operator,
to make sure that when `os.platform` is empty we get
the default ABI value, and thus fixes build on riscv64.
[1] https://www.boost.org/doc/libs/1_83_0/tools/build/doc/html/index.html#jam.language.flow_of_control
Fixes: d039c8e4da
On C++11 static local variables are initialized in thread-safe manner, but even on C++03 it should not be a problem because in our case variables are of trivial types, which means double initialization is not an issue, and they are initialized with the same value in every thread.
This fixes fcontext on my PowerBook G4 running Void Linux
ppc-musl-20190901, NetBSD/macppc 8.1, or OpenBSD/macppc 6.6-current,
all with g++. These systems use fcontext for *ppc32_sysv_elf*
(PowerPC 32-bit System V ELF). The assembly code was wrong for BSD
and crashing on Linux musl.
Linux returns a transfer_t in memory (through a hidden pointer in R3),
but other systems (at least NetBSD and OpenBSD) return a transfer_t in
registers R3:R4. jump_fcontext() and ontop_fcontext() were always
using the hidden pointer. Add checks for `#ifdef__linux__`; start
using R3:R4 on other systems.
make_fcontext() was calling _exit(0) through the insecure BSS PLT.
Set R30 to use the secure PLT. This prevents a crash when musl's
ld.so loads the executable; musl seems to require the secure PLT.
Fix ontop_fcontext() to restore the hidden pointer on Linux. It was
passing the wrong context's hidden pointer to the ontop-function fn(),
so fn() returned a transfer_t to the wrong stack. When fn() was
context_exit() in <boost/context/continuation_fcontext.hpp>, it freed
the old stack, then returned `transfer_t{ nullptr, nullptr }` to free
memory. This crashed on Linux musl.
Now that ontop_fcontext() restores the hidden pointer, it must stop
abusing the same pointer to pass a transfer_t argument to fn(). Add a
new ontop_fcontext_tail() in C++, which takes arguments in registers
and allocates a transfer_t. The code is in C++ so it can free the
transfer_t argument if fn() throws a C++ exception.
Rearrange the context frame to shrink it from 244 to 240 bytes. This
fixes the stack alignment: the ABI requires R1 % 16 == 0, and
make_fcontext() respects this, but jump_fcontext() was adding 244 to
R1, so the new context ran with a misaligned stack (244 % 16 == 4).
Remove R13 from the context frame, so new contexts stop loading R13
with garbage. The ABI uses R13 to point to the executable's small
data, so R13 should have the same value in every context.
Add the backchain to the context frame; make room by moving LR to the
caller's frame. Order CR, R14 to R31, F14 to F31 at the frame's end,
as is typical for this ABI. Provide 8-byte alignment for FPSCR and
F14 to F31, to avoid a misalignment penalty.
