As per discussion with Ulrich Weigand, document the restrictions
on the code within ffi_call_int as we simultaneously prepare
stack frames for ffi_call_SYSV and the target function.
It's silly to stick to esa/390 features when the compiler won't.
Detect when brasl and larl are used by the compiler and then use
them in the assembly.
Since libffi currently doesn't allow empty structures, libgo
currently maps them to ffi_type_void. Given that we'll abort
on this case, handle it gracefully.
Unlike ffi_closure_LINUX64, this entry point is called normally,
so we already have the TOC in R2 and the closure in R11.
* powerpc/linux64_closure.S (ffi_closure_LINUX64): Remove a
register dependency chain.
(ffi_go_closure_linux64): Don't load r11 or r2.
Plus .cfi async unwind info, rearrangement of ffi_call_linux64 and
ffi_call_SYSV function params to avoid register copies, tweaks to
trampolines.
* src/powerpc/ffitarget.h (FFI_GO_CLOSURES): Define.
* src/powerpc/ffi.c (ffi_call_int): New function with extra
closure param, and args rearranged on ffi_call_linux64 and
ffi_call_SYSV calls, extracted from ..
(ffi_call): ..here.
(ffi_call_go, ffi_prep_go_closure): New functions.
* src/powerpc/ffi_linux64.c (ffi_prep_closure_loc_linux64): Make
hidden. Only flush insn part of ELFv2 trampoline. Don't shuffle
ELFv1 trampoline.
(ffi_closure_helper_LINUX64): Replace closure param with cif, fun,
user_data params.
* src/powerpc/ffi_powerpc.h (ffi_go_closure_sysv): Declare.
(ffi_go_closure_linux64): Declare.
(ffi_call_SYSV, fi_call_LINUX64): Update.
(ffi_prep_closure_loc_sysv, ffi_prep_closure_loc_linux64): Declare.
(ffi_closure_helper_SYSV, ffi_closure_helper_LINUX64): Update.
* src/powerpc/ffi_sysv.c (ASM_NEEDS_REGISTERS): Increase to 6.
(ffi_prep_closure_loc_sysv): Use bcl in trampoline, put data words
last, flush just the insn part.
(ffi_closure_helper_SYSV): Replace closure param with cif, fun and
user_data params.
* src/powerpc/linux64.S (ffi_call_LINUX64): Replace hand-written
.eh_frame with .cfi directives. Adjust for changed param order.
Pass extra "closure" param to user function in static chain. Add
.cfi directives to describe epilogue. Don't provide traceback
table for ELFv2 or _CALL_LINUX.
* src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Replace
hand-written .eh_frame with .cfi directives. Adjust for changed
ffi_closure_helper_LINUX64 params. Add .cfi directives to
describe epilogue. Don't provide traceback table for ELFv2 or
_CALL_LINUX.
(ffi_go_closure_linux64): New function.
* src/powerpc/sysv.S: Remove redundant .globl ffi_prep_args_SYSV.
(ffi_call_SYSV): Make hidden. Replace hand-written .eh_frame with
.cfi directives. Adjust for changed params. Pass extra "closure"
param to user function in static chain. Add .cfi directives to
describe epilogue.
* src/powerpc/ppc_closure.S (ffi_closure_SYSV): Make hidden.
Replace hand-written .eh_frame with .cfi directives. Adjust for
changed ffi_closure_helper_SYSV params. Add .cfi directives to
describe epilogue. Don't just use nops in the dead __NO_FPRS__
epilogues.
(ffi_go_closure_sysv): New function.
This is a port of
http://gcc.gnu.org/viewcvs?rev=207763&root=gcc&view=rev
aka GCC PR libffi/60073, to the rewritten Sparc codebase.
Supposedly, we should have seen failures with the existing
libffi.call/cls_double_va.c testcase, but I hadn't.
Perhaps a gcc newer than 4.6.3 is required to see that...
The original code, removed in the "rewrite" patch, was incorrect for
large structures, and required dynamic allocation of a trampoline on
every ffi_call.
Instead, allocate a 4k entry table of all possible structure returns.
The table is 80k, but is read-only and dynamically paged, which ought
to be better than allocating the trampoline.
This is difficult to test with gcc. One can only use -O0 at present.
See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63668.
It's impossible to call between v8 and v9 ABIs, because of the stack bias
in the v9 ABI. So let's not pretend it's just not implemented yet. Split
the v9 code out to a separate file.
The register windows prevent ffi_call from setting up the entire stack
frame the assembly, but we needn't make an indirect call back to prep_args.
Assembly to use local labels, .type annotation, hidden annotation.
I do retain the _prefix for the symbols, but given that it wasn't
done consistently across all symbols, I doubt it's actually needed.
Do not modify the ffi_type. Rearrange the tests so that we
quickly eliminate structures that cannot match. Return an
encoded value of element count and base type.
Unties the backend from changes to FFI_TYPE_* constants, and allows
compilation to succeed after the addition of FFI_TYPE_COMPLEX.
Delete the hand-written unwind info.
Richard Henderson
Nobuyoshi Nakada
Richard Henderson
Dominik Vogt
Anthony Green
James Greenhalgh
Alan Modra
Gregory Pakosz