The ELFv1 ABI says: "Single precision floating point values are mapped
to the second word in a single doubleword" and also "Floating point
registers f1 through f13 are used consecutively to pass up to 13
floating point values, one member aggregates passed by value
containing a floating point value, and to pass complex floating point
values".
libffi wasn't expecting float args in the second word, and wasn't
passing one member aggregates in fp registers. This patch fixes those
problems, making use of the existing ELFv2 homogeneous aggregate
support since a one element fp struct is a special case of an
homogeneous aggregate.
I've also set a flag when returning pointers that might be used one
day. This is just a tidy since the ppc64 assembly support code
currently doesn't test FLAG_RETURNS_64BITS for integer types..
* src/powerpc/ffi_linux64.c (discover_homogeneous_aggregate):
Compile for ELFv1 too, handling single element aggregates.
(ffi_prep_cif_linux64_core): Call discover_homogeneous_aggregate
for ELFv1. Set FLAG_RETURNS_64BITS for FFI_TYPE_POINTER return.
(ffi_prep_args64): Call discover_homogeneous_aggregate for ELFv1,
and handle single element structs containing float or double
as if the element wasn't wrapped in a struct. Store floats in
second word of doubleword slot when big-endian.
(ffi_closure_helper_LINUX64): Similarly.
Issue #70 pointed out that at least one compiler didn't like:
return ffi_call_efi64(cif, fn, rvalue, avalue);
... where the return type is "void". This patch splits the statement
into two.
I also noticed that ffi_call_go here seems to do a double call. I
suspect a "return" is missing here, so this patch adds it as well.
On CentOS 7.4, configure script generated by autogen.sh will output the
following values when compiling GCC source code
toolexecdir='NONE/$(target_alias)'
toolexeclibdir='NONE/$(target_alias)/lib'
and cause build error
... -o libffi.la -rpath NONE/riscv64-bs-linux-gnu/lib ...
checking for shl_load in -ldld... libtool: link: only absolute run-paths are allowed
* mips: simplify closure #defines
This commit should have no visible effect.
* mips: add special handling of variadic functions
MIPS requires special handling of variadic functions which pass floating
point arguments:
* In the o32 ABI, all float arguments are passed in integer registers.
* In the n32/n64 ABIs, float arguments after the ellipsis are passed in
integer registers.
Implement this in libffi. To support this in n32/n64 closures, we need to add
a new mips_nfixedargs field to ffi_cif which will break the libffi ABI.
This fixes the libffi.call/cls_longdouble_va.c test which was failing on
64-bit MIPS.
* mips: align argn for all 64-bit types in o32 closure handler
Ensure that argn is pre-aligned for all 64-bit argument types (including
doubles) and not just integer types.
This fixes closures of the form "f(float, double, <some integer args>)".
Previously the first integer argument would be read from a2 which is garbage
at this point (the float arguments have already "consumed" a0-a3). After
this commit, argn is correctly padded between the "float" and "double"
arguments so that the first integer argument is read from the stack.
Fixes "double f(float,double,int)" test in #371
* mips: do not read from floating point register if returning a struct
In the o32 ABI, the pointer passed in a0 used to return structures
indirectly is treated as the first argument for argument allocation purposes.
This means that it should inhibit floating point registers the same way that
other integer arguments do.
Fixes "Double f(float,Double,double)" test in #371
* mips: fix pointer cast warnings
Fix two pointer cast warnings when compiled on 64-bit mips by casting
through uintptr_t.
Fixes mips64el part of #404
* Fix msvcc dll build by adding dllexport decorations to all API declarations
* Fix appveyor build for VS 2013
Use the new -DFFI_BUILDING_DLL for producing a working DLL. Update the
msvcc.sh wrapper script to successfully compile the testsuite files.
* MSVC build: suppress warnings in testsuite
* fix testsuite on appveyor