2008-02-26 Anthony Green <green@redhat.com>
Thomas Heller <theller@ctypes.org>
* src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C
comment.
2008-02-03 Timothy Wall <twall@users.sf.net>
* src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return
offset based on code pointer, not data pointer.
2008-01-31 Timothy Wall <twall@users.sf.net>
* testsuite/libffi.call/closure_stdcall.c: Add test for stdcall
closures.
* src/x86/ffitarget.h: Increase size of trampoline for stdcall
closures.
* src/x86/win32.S: Add assembly for stdcall closure.
* src/x86/ffi.c: Initialize stdcall closure trampoline.
From-SVN: r148174
2008-08-25 Andreas Tobler <a.tobler@schweiz.org>
* src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and
FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum.
Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT.
Adjust copyright notice.
* src/powerpc/ffi.c: Add two new flags to indicate if we have one
register or two register to use for FFI_SYSV structs.
(ffi_prep_cif_machdep): Pass the right register flag introduced above.
(ffi_closure_helper_SYSV): Fix the return type for
FFI_SYSV_TYPE_SMALL_STRUCT. Comment.
Adjust copyright notice.
From-SVN: r139567
2008-01-05 Andreas Tobler <a.tobler@schweiz.org>
PR testsuite/32843
* src/x86/ffi.c (ffi_prep_cif_machdep): Add code for
signed/unsigned int8/16 for X86_DARWIN.
Updated copyright info.
Handle one and two byte structs with special cif->flags.
* src/x86/ffitarget.h: Add special types for one and two byte structs.
Updated copyright info.
* src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like
sysv.S
Remove code to pop args from the stack after call.
Special-case signed/unsigned for int8/16, one and two byte structs.
(ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8,
FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32,
FFI_TYPE_SINT32.
Updated copyright info.
From-SVN: r131343
2007-12-08 David Daney <ddaney@avtrex.com>
* src/mips/n32.S (ffi_call_N32): Replace dadd with ADDU, dsub with
SUBU, add with ADDU and use smaller code sequences.
From-SVN: r130711
2007-12-06 David Daney <ddaney@avtrex.com>
* include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already
defined.
(ffi_java_raw): New typedef.
(ffi_java_raw_call, ffi_java_ptrarray_to_raw,
ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to
ffi_java_raw.
(ffi_java_raw_closure) : Same.
(ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change
parameter types.
* src/java_raw_api.c (ffi_java_raw_size): Replace FFI_SIZEOF_ARG with
FFI_SIZEOF_JAVA_RAW.
(ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw.
Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use
sizeof(ffi_java_raw) for alignment calculations.
(ffi_java_ptrarray_to_raw): Same.
(ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER
if FFI_SIZEOF_JAVA_RAW == 4.
(ffi_java_raw_to_rvalue): Same.
(ffi_java_raw_call): Change type of raw to ffi_java_raw.
(ffi_java_translate_args): Same.
(ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change
parameter types.
* src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI.
2007-12-06 David Daney <ddaney@avtrex.com>
* interpret.cc: Replace ffi_raw with INTERP_FFI_RAW_TYPE throughout.
(ncode_closure, ffi_closure_fun): Define versions for
non-FFI_NATIVE_RAW_API case.
* include/java-interp.h (INTERP_FFI_RAW_TYPE): Define and use to
replace ffi_raw throughout.
* jni.cc, interpret-run.cc: Replace ffi_raw with INTERP_FFI_RAW_TYPE
throughout.
From-SVN: r130660
2007-12-01 Andreas Tobler <a.tobler@schweiz.org>
PR libffi/31937
* src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT.
Add local FFI_TYPE_UINT128 to handle soft-float long-double-128.
* src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and
set the NUM_FPR_ARG_REGISTERS according to.
Add support for potential soft-float support under hard-float
architecture.
(ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of
FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according
to the FFI_LINUX_SOFT_FLOAT ABI.
(ffi_prep_cif_machdep): Likewise.
(ffi_closure_helper_SYSV): Likewise.
* src/powerpc/ppc_closure.S: Make sure not to store float/double
on archs where __NO_FPRS__ is true.
Add FFI_TYPE_UINT128 support.
* src/powerpc/sysv.S: Add support for soft-float long-double-128.
Adjust copyright notice.
From-SVN: r130559
2007-11-25 Andreas Tobler <a.tobler@schweiz.org>
* src/closures.c: Move defintion of MAYBE_UNUSED from here to ...
* include/ffi_common.h: ... here.
Update copyright.
From-SVN: r130413
2007-11-17 Andreas Tobler <a.tobler@schweiz.org>
* src/powerpc/sysv.S: Load correct cr to compare if we have long double.
* src/powerpc/linux64.S: Likewise.
* src/powerpc/ffi.c: Add a comment to show which part goes into cr6.
* testsuite/libffi.call/return_ldl.c: New test.
From-SVN: r130247
2007-09-04 <aph@redhat.com>
* src/arm/sysv.S (UNWIND): New.
(Whole file): Conditionally compile unwinder directives.
* src/arm/sysv.S: Add unwinder directives.
* src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes.
Only treat r0 as a struct address if we're actually returning a
struct by address.
Only copy the bytes that are actually within a struct.
(ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes
is returned in r0, not passed by address.
(ffi_call): Allocate a word-sized temporary for the case where
a composite is returned in r0.
(ffi_prep_incoming_args_SYSV): Align as necessary.
2007-08-05 Steven Newbury <s_j_newbury@yahoo.co.uk>
* src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of
directly using the sys_cacheflush syscall.
2007-07-27 Andrew Haley <aph@redhat.com>
* src/arm/sysv.S (ffi_closure_SYSV): Add soft-float.
From-SVN: r128093
PR libffi/28313
* configure.ac: Don't treat mips64 as a special case.
* Makefile.am (nodist_libffi_la_SOURCES): Add n32.S.
* configure: Regenerate
* Makefile.in: Ditto.
* fficonfig.h.in: Ditto.
* src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent.
(LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros.
(FFI_DEFAULT_ABI): Set for n64 case.
(FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases.
* src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE.
(ffi_closure_N32): New function.
(.eh_frame): New section
* src/mips/o32.S: Clean up comments.
(ffi_closure_O32): Pass ffi_closure parameter in $12.
* src/mips/ffi.c: Use FFI_MIPS_N32 instead of
_MIPS_SIM == _ABIN32 throughout.
(FFI_MIPS_STOP_HERE): New, use in place of
ffi_stop_here.
(ffi_prep_args): Use unsigned long to hold pointer values. Rewrite
to support n32/n64 ABIs.
(calc_n32_struct_flags): Rewrite.
(calc_n32_return_struct_flags): Remove unused variable. Reverse
position of flag bits.
(ffi_prep_cif_machdep): Rewrite n32 portion.
(ffi_call): Enable for n64. Add special handling for small structure
return values.
(ffi_prep_closure_loc): Add n32 and n64 support.
(ffi_closure_mips_inner_O32): Add cast to silence warning.
(copy_struct_N32, ffi_closure_mips_inner_N32): New functions.
From-SVN: r127336
2007-07-30 Andrew Haley <aph@redhat.com>
PR testsuite/32843
* src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for
signed/unsigned int8/16.
* src/x86/sysv.S (ffi_call_SYSV): Rewrite to:
Use a jump table.
Remove code to pop args from the stack after call.
Special-case signed/unsigned int8/16.
* testsuite/libffi.call/return_sc.c (main): Revert.
From-SVN: r127093
* configure.ac: Tidy target case.
(HAVE_LONG_DOUBLE): Allow the target to override.
* configure: Regenerate.
* include/ffi.h.in: Don't define ffi_type_foo if
LIBFFI_HIDE_BASIC_TYPES is defined.
(ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define
to ffi_type_double.
* types.c (LIBFFI_HIDE_BASIC_TYPES): Define.
(FFI_TYPEDEF, ffi_type_void): Mark the data const.
(ffi_type_longdouble): Special case for Alpha. Don't define
if long double == double.
* src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value.
(ffi_prep_cif_machdep): Handle it as the 128-bit type.
(ffi_call, ffi_closure_osf_inner): Likewise.
(ffi_closure_osf_inner): Likewise. Mark hidden.
(ffi_call_osf, ffi_closure_osf): Mark hidden.
* src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition.
* src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden.
(load_table): Handle 128-bit long double.
* testsuite/libffi.call/float4.c: Add -mieee for alpha.
From-SVN: r123622
* src/closures.c: Include sys/statfs.h.
(_GNU_SOURCE): Define on Linux.
(FFI_MMAP_EXEC_SELINUX): Define.
(selinux_enabled): New variable.
(selinux_enabled_check): New function.
(is_selinux_enabled): Define.
(dlmmap): Use it.
* configure.ac (NO_EXECUTE_PERMISSION): Set by default.
* configure: Rebuilt.
From-SVN: r123457
2007-02-13 Andreas Krebbel <krebbel1@de.ibm.com>
* src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep,
ffi_closure_helper_SYSV): Add long double handling.
From-SVN: r121879
2006-12-14 Andreas Tobler <a.tobler@schweiz.org>
* configure.ac: Add TARGET for x86_64-*-darwin*.
* Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources
for X86_DARWIN.
* src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*.
* src/x86/darwin64.S: New file for x86_64-*-darwin* support.
* configure: Regenerate.
* Makefile.in: Regenerate.
* include/Makefile.in: Regenerate.
* testsuite/Makefile.in: Regenerate.
* testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for
ffi_call only.
From-SVN: r119856
2006-04-05 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
Andreas Tobler <a.tobler@schweiz.ch>
* Makefile.am: Add PA_HPUX port.
* Makefile.in: Regenerate.
* include/Makefile.in: Likewise.
* testsuite/Makefile.in: Likewise.
* configure.ac: Add PA_HPUX rules.
* configure: Regenerate.
* src/pa/ffitarget.h: Rename linux target to PA_LINUX.
Add PA_HPUX and PA64_HPUX.
Rename FFI_LINUX ABI to FFI_PA32 ABI.
(FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets.
(FFI_TYPE_SMALL_STRUCT2): Define.
(FFI_TYPE_SMALL_STRUCT4): Likewise.
(FFI_TYPE_SMALL_STRUCT8): Likewise.
(FFI_TYPE_SMALL_STRUCT3): Redefine.
(FFI_TYPE_SMALL_STRUCT5): Likewise.
(FFI_TYPE_SMALL_STRUCT6): Likewise.
(FFI_TYPE_SMALL_STRUCT7): Likewise.
* src/pa/ffi.c (ROUND_DOWN): Delete.
(fldw, fstw, fldd, fstd): Use '__asm__'.
(ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2,
FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8.
(ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment.
Simplify incrementing of stack slot variable. Change type of local
'n' to unsigned int.
(ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long
double on PA_HPUX.
(ffi_prep_cif_machdep): Likewise.
(ffi_call): Likewise.
(ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change
return type to ffi_status. Simplify incrementing of stack slot
variable. Only copy floating point argument registers when PA_LINUX
is true. Reformat debug statement.
Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and
FFI_TYPE_SMALL_STRUCT8.
(ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to
declaration.
(ffi_prep_closure): Make linux trampoline conditional on PA_LINUX.
Add nops to cache flush. Add trampoline for PA_HPUX.
* src/pa/hpux32.S: New file.
* src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename
ffi_prep_args_LINUX to ffi_prep_args_pa32.
Localize labels. Add support for 2, 4 and 8-byte small structs. Handle
unaligned destinations in 3, 5, 6 and 7-byte small structs. Order
argument type checks so that common argument types appear first.
(ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename
ffi_closure_inner_LINUX to ffi_closure_inner_pa32.
Co-Authored-By: Andreas Tobler <a.tobler@schweiz.ch>
From-SVN: r112719
* src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX. Default
for 32-bit using IBM extended double format. Fix FFI_LAST_ABI.
* src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of
FFI_TYPE_LONGDOUBLE.
(ffi_prep_args64): Assert using IBM extended double.
(ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type.
Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args.
(ffi_call): Handle FFI_LINUX.
(ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs
gpr3 return pointer as for struct return. Handle FFI_LINUX
FFI_TYPE_LONGDOUBLE return and args. Don't increment "nf"
unnecessarily.
* src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2
for FFI_TYPE_LONGDOUBLE. Move epilogue insns into case table.
Don't use r6 as pointer to results, instead use sp offset. Don't
make a special call to load lr with case table address, instead
use offset from previous call.
* src/powerpc/sysv.S (ffi_call_SYSV): Save long double return.
* src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double
return.
From-SVN: r112340
* src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable
and cast integer to void * if needed. Update the pointer to
the FP register saved area correctly.
From-SVN: r111225
* src/powerpc/ffi.c: Formatting.
(ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions.
(ffi_prep_args64): Likewise.
From-SVN: r107963
* testsuite/libffi.call/float1.c (value_type): New typedef.
(CANARY): New define.
(main): Check for result buffer overflow.
* src/powerpc/linux64.S: Handle linux64 long double returns.
* src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant.
(ffi_prep_cif_machdep): Handle linux64 long double returns.
From-SVN: r104660
* src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use
DW_CFA_offset_extended_sf rather than
DW_CFA_GNU_negative_offset_extended.
* src/powerpc/sysv.S (ffi_call_SYSV): Likewise.
From-SVN: r102900
* src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly
on sh3.
(ffi_closure_SYSV): Change the stack layout for sh3 struct argument.
* src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is
partially on register.
(ffi_closure_helper_SYSV): Likewise.
(ffi_prep_cif_machdep): Don't set too many cif->flags.
From-SVN: r102298
* src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in
bit 11 of flags.
(ffi_call): Mask return type field. Pass ssecount to ffi_call_unix64.
(ffi_prep_closure): Set carry bit if sse-used flag set.
* src/x86/unix64.S (ffi_call_unix64): Add ssecount argument.
Only load sse registers if ssecount non-zero.
(ffi_closure_unix64): Only save sse registers if carry set on entry.
From-SVN: r99257
* src/types.c (FFI_AGGREGATE_TYPEDEF): Remove.
(FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF. Replace size and
offset parameters with a type parameter; deduce size and structure
alignment. Update all users.
From-SVN: r92775
* src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead
of jal. Use an absolute encoding for the frame information.
* configure.host: Enable the interpreter for mips*-*-linux*.
From-SVN: r91116
* src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers
on a 8-byte boundary.
* src/sparc/v8.S (ffi_closure_v8): Reserve frame space for
arguments.
From-SVN: r91022
* src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return
long long values. Round stack allocation to a multiple of 8 bytes
for ATPCS compatibility.
* src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register
names. Handle returning long long types. Add Thumb and interworking
support. Improve soft-float code.
From-SVN: r89681
* src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for
the function returning a structure pointed with R2.
* src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to
the structure return value if T bit set. Emit position
independent code and EH data if PIC.
From-SVN: r89254
2004-09-02 Andreas Tobler <a.tobler@schweiz.ch>
* src/powerpc/ffi_darwin.c: Add flag for longdouble return values.
(ffi_prep_args): Handle longdouble arguments.
(ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for
longdouble.
(ffi_closure_helper_DARWIN): Add closure handling for longdouble.
* src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble
values.
* src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise.
* src/types.c: Defined longdouble size and alignment for darwin.
From-SVN: r86992
2004-08-30 Anthony Green <green@redhat.com>
* Makefile.am: Add frv support.
* Makefile.in, testsuite/Makefile.in: Rebuilt.
* configure.ac: Read configure.host.
* configure, aclocal.m4: Rebuilt.
* configure.host: New file. frv-elf needs libgloss.
* include/ffi.h.in: Force ffi_closure to have a nice big (8)
alignment. This is needed to frv and should harm the others.
* include/ffi_common.h (ALIGN_DOWN): New macro.
* src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files.
From-SVN: r86779
2004-08-14 Casey Marshall <csm@gnu.org>
* src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to
contain `FFI_TYPE_UINT64' as return type for any 64-bit
integer (O32 ABI only).
(ffi_prep_closure): new function.
(ffi_closure_mips_inner_O32): new function.
* src/mips/ffitarget.h: Define `FFI_CLOSURES' and
`FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32.
* src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return
64 bit integers correctly.
(ffi_closure_O32): new function.
Added DWARF-2 unwind info for both functions.
From-SVN: r86019
* target.h (struct gcc_target): New target hook, unwind_label.
* target-def.h (TARGET_ASM_EMIT_UNWIND_LABEL): New hook.
* output.h (default_emit_unwind_label): New function.
* default.h (TARGET_EXPLICIT_INSTANTIATIONS_ONE_ONLY): New macro.
(TARGET_USES_WEAK_UNWIND_INFO): New target macro.
(TARGET_SUPPORTS_HIDDEN): New target macro.
* dwarf2out.c (struct dw_fde_struct): Add field for function decl
that corresponds to this FDE.
(FRAME_BEGIN_LABEL): Allow target to override default label.
(output_call_frame_info): If FDEs are linknonce, then use extra
indirection for FDE encoding, output a label for each FDE, and
output an empty label for each function without an FDE.
(dwarf2out_begin_prologue): Set up decl field when creating an FDE.
* varasm.c (globalize_decl): Call ASM_MAKE_LABEL_LINKONCE for
decls with DECL_ONE_ONLY set, if that macro is defined.
(make_decl_one_only): Don't use DECL_COMMON if we're compiling
for a SUPPORTS_ONE_ONLY target.
* config/darwin-protos.h (darwin_unique_section): Declare.
(darwin_asm_named_section): Likewise.
(darwin_section_type_flags): Likewise.
(darwin_non_lazy_pcrel): Likewise.
(darwin_emit_unwind_label): Likewise.
(darwin_make_decl_one_only): Likewise.
* config/darwin.c (machopic_finish): Get rid of tweak that
eliminate stubs for symbols that are defined.
(darwin_encode_section_info): Don't treat weak functions as defined.
(darwin_make_decl_one_only): Define.
(darwin_asm_named_section): Likewise.
(darwin_section_type_flags): Likewise.
(darwin_unique_section): Likewise.
(darwin_emit_unwind_label): Likewise.
(darwin_non_lazy_pcrel): Likewise.
(darwin_asm_output_dwarf_delta): Difference between two labels is
local only if both labels are local.
* config/darwin.h (MAKE_DECL_ONE_ONLY): Define.
(ASM_MAKE_LABEL_LINKONCE): Likewise.
(TARGET_SUPPORTS_HIDDEN): Likewise.
(TARGET_USES_WEAK_UNWIND_INFO): Likewise.
(TARGET_EXPLICIT_INSTANTIATIONS_ONE_ONLY): Likewise.
(FRAME_BEGIN_LABEL): Likewise.
(ASM_DECLARE_OBJECT_NAME): Make references to weak symbols indirect.
(ASM_DECLARE_FUNCTION_NAME): Likewise.
(darwin_eh_frame_section): Give __eh_frame section the coalesced flag.
(TARGET_ASM_UNIQUE_SECTION): Define.
(EH_FRAME_SECTION_NAME): Define.
(EH_FRAME_SECTION_ATTR): Likewise.
(ASM_MAYBE_OUTPUT_ENCODED_ADDR_RTX): Likewise.
(TARGET_ASM_NAMED_SECTION): Likewise.
(TARGET_SECTION_TYPE_FLAGS): Likewise.
* doc/tm.texi: Document TARGET_USES_WEAK_UNWIND_INFO,
TARGET_EXPLICIT_INSTANTIATIONS_ONE_ONLY, TARGET_SUPPORTS_HIDDEN,
TARGET_ASM_EMIT_UNWIND_LABEL.
* cp/decl2.c (maybe_make_one_only): Look at
TARGET_EXPLICIT_INSTANTIATION_ONE_ONLY when deciding whether
to make an explicit instantiation weak.
* cp/method.c (use_thunk): Make sure we call comdat_linkage
when appropriate.
* cp/pt.c (do_type_instantiation): On systems where weak symbols
don't go in a static archive's TOC, explicit instantiation of a
class must imply *explicit* instantiation of its memeber.
From-SVN: r79394
* src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point
arguments from fp registers only for the first 8 parameter slots.
Don't convert a float parameter when passed in memory.
From-SVN: r79323
* src/powerpc/ffi.c (ffi_prep_cif_machdep <FFI_LINUX64>): Correct
long double function return and long double arg handling.
(ffi_closure_helper_LINUX64): Formatting. Delete unused "ng" var.
Use "end_pfr" instead of "nf". Correct long double handling.
Localise "temp".
* src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double
return value.
* src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate
space for long double return value. Adjust stack frame and offsets.
Load f2 long double return.
From-SVN: r77481
* src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array
when the structure return address is passed in %o0.
(ffi_V9_return_struct): Rename into ffi_v9_layout_struct.
(ffi_v9_layout_struct): Align the field following a nested structure
on a word boundary. Use memmove instead of memcpy.
(ffi_call): Update call to ffi_V9_return_struct.
(ffi_prep_closure): Define 'ctx' only for V8.
(ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8
and ffi_closure_sparc_inner_v9.
(ffi_closure_sparc_inner_v8): Return long doubles by reference.
Always skip the structure return address. For structures and long
doubles, copy the argument directly.
(ffi_closure_sparc_inner_v9): Skip the structure return address only
if required. Shift the maximum floating-point slot accordingly. For
big structures, copy the argument directly; otherwise, left-justify the
argument and call ffi_v9_layout_struct to lay out the structure on
the stack.
* src/sparc/v8.S: Undef STACKFRAME before defining it.
(ffi_closure_v8): Pass the structure return address. Update call to
ffi_closure_sparc_inner_v8. Short-circuit FFI_TYPE_INT handling.
Skip the 'unimp' insn when returning long doubles and structures.
* src/sparc/v9.S: Undef STACKFRAME before defining it.
(ffi_closure_v9): Increase the frame size by 2 words. Short-circuit
FFI_TYPE_INT handling. Load structures both in integers and
floating-point registers on return.
* README: Update status of the SPARC port.
From-SVN: r76543
2003-10-22 David Daney <ddaney@avtrex.com>
* src/mips/ffitarget.h: Replace undefined UINT32 and friends with
__attribute__((__mode__(__SI__))) and friends.
From-SVN: r72802
2003-10-21 Andreas Tobler <a.tobler@schweiz.ch>
* configure.in: AC_LINK_FILES(ffitarget.h).
* configure: Regenerate.
* Makefile.in: Likewise.
* include/Makefile.in: Likewise.
* testsuite/Makefile.in: Likewise.
* fficonfig.h.in: Likewise.
2003-10-21 Paolo Bonzini <bonzini@gnu.org>
Richard Henderson <rth@redhat.com>
Avoid that ffi.h includes fficonfig.h.
* Makefile.am (EXTRA_DIST): Include ffitarget.h files
(TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
(TARGET_SRC_MIPS_SGI): Removed.
(MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX.
(MIPS_SGI): Removed.
(CLEANFILES): Removed.
(mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New
targets.
* acconfig.h: Removed.
* configure.in: Compute sizeofs only for double and long double.
Use them to define and subst HAVE_LONG_DOUBLE. Include comments
into AC_DEFINE instead of using acconfig.h. Create
include/ffitarget.h instead of include/fficonfig.h. Rename
MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree.
AC_DEFINE EH_FRAME_FLAGS.
* include/Makefile.am (DISTCLEANFILES): New automake macro.
(hack_DATA): Add ffitarget.h.
* include/ffi.h.in: Remove all system specific definitions.
Declare raw API even if it is not installed, why bother?
Use limits.h instead of SIZEOF_* to define ffi_type_*. Do
not define EH_FRAME_FLAGS, it is in fficonfig.h now. Include
ffitarget.h instead of fficonfig.h. Remove ALIGN macro.
(UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead.
* include/ffi_common.h (bool): Do not define.
(ffi_assert): Accept failed assertion.
(ffi_type_test): Return void and accept file/line.
(FFI_ASSERT): Pass stringized failed assertion.
(FFI_ASSERT_AT): New macro.
(FFI_ASSERT_VALID_TYPE): New macro.
(UINT8, SINT8, UINT16, SINT16, UINT32, SINT32,
UINT64, SINT64): Define here with gcc's __attribute__ macro
instead of in ffi.h
(FLOAT32, ALIGN): Define here instead of in ffi.h
* include/ffi-mips.h: Removed. Its content moved to
src/mips/ffitarget.h after separating assembly and C sections.
* src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c
src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c,
src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S,
src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c:
SIZEOF_ARG -> FFI_SIZEOF_ARG.
* src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+).
* src/debug.c (ffi_assert): Accept stringized failed assertion.
(ffi_type_test): Rewritten.
* src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call
FFI_ASSERT_VALID_TYPE.
* src/alpha/ffitarget.h, src/arm/ffitarget.h,
src/ia64/ffitarget.h, src/m68k/ffitarget.h,
src/mips/ffitarget.h, src/powerpc/ffitarget.h,
src/s390/ffitarget.h, src/sh/ffitarget.h,
src/sh64/ffitarget.h, src/sparc/ffitarget.h,
src/x86/ffitarget.h: New files.
* src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S,
src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S,
src/powerpc/aix.S, src/powerpc/darwin.S,
src/powerpc/ffi_darwin.c, src/powerpc/linux64.S,
src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S,
src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S,
src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S,
src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S:
include fficonfig.h
From-SVN: r72766
2003-09-18 Andreas Tobler <a.tobler@schweiz.ch>
David Edelsohn <edelsohn@gnu.org>
* src/types.c (double): Add AIX and Darwin to the right TYPEDEF.
* src/powerpc/aix_closure.S: Remove the pointer to the outgoing
parameter stack.
* src/powerpc/darwin_closure.S: Likewise.
* src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures
according to the Darwin/AIX ABI.
(ffi_prep_cif_machdep): Likewise.
(ffi_closure_helper_DARWIN): Likewise.
Remove the outgoing parameter stack logic. Simplify the evaluation
of the different CASE types.
(ffi_prep_clousure): Avoid the casts on lvalues. Change the branch
statement in the trampoline code.
Co-Authored-By: David Edelsohn <edelsohn@gnu.org>
From-SVN: r71530
* src/sh/ffi.c (ffi_prep_args): Take account into the alignement
for the register size.
(ffi_closure_helper_SYSV): Handle the structure return value
address correctly.
(ffi_closure_helper_SYSV): Return the appropriate type when
the registers are used for the structure return value.
* src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for
the 64-bit return value. Update copyright years.
From-SVN: r71521
* src/types.c (double, longdouble): Merge identical SH and ARM
typedefs, and add POWERPC64.
* src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for
struct split over gpr and rest.
(ffi_prep_cif_machdep): Correct intarg_count for structures.
* src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets.
From-SVN: r71295
2003-08-30 Andreas Tobler <a.tobler@schweiz.ch>
* src/java_raw_api.c: Include <stdlib.h> to fix compiler warning
about implicit declaration of abort().
From-SVN: r71071
* src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc
structure passing.
(ffi_closure_helper_LINUX64): Likewise.
* src/powerpc/linux64.S: Remove code writing to parm save area.
* src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return
address in lr from ffi_closure_helper_LINUX64 call to calculate
table address. Optimize function tail.
From-SVN: r70050
2003-07-28 Andreas Tobler <a.tobler@schweiz.ch>
* src/sparc/ffi.c: Handle all floating point registers.
* src/sparc/v9.S: Likewise. Fixes second part of PR target/11410.
From-SVN: r69951
* configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section
should be read-only.
* configure: Rebuilt.
* fficonfig.h.in: Rebuilt.
* include/ffi.h.in (EH_FRAME_FLAGS): Define.
* src/alpha/osf.S: Use EH_FRAME_FLAGS.
* src/powerpc/linux64.S: Likewise.
* src/powerpc/linux64_closure.S: Likewise. Include ffi.h.
* src/powerpc/sysv.S: Use EH_FRAME_FLAGS. Use pcrel encoding
if -fpic/-fPIC/-mrelocatable.
* src/powerpc/powerpc_closure.S: Likewise.
* src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include
#write in .eh_frame flags.
* src/sparc/v9.S: Likewise.
* src/x86/unix64.S: Use EH_FRAME_FLAGS.
* src/x86/sysv.S: Likewise. Use pcrel encoding if -fpic/-fPIC.
* src/s390/sysv.S: Use EH_FRAME_FLAGS. Include ffi.h.
From-SVN: r66887
* configure.in (powerpc64*-*-linux*): Remove.
* configure: Rebuilt.
libffi/
* include/ffi.h.in (POWERPC64): Define if 64-bit.
(enum ffi_abi): Add FFI_LINUX64 on POWERPC.
Make it the default on POWERPC64.
(FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64.
* configure.in: Change powerpc-*-linux* into powerpc*-*-linux*.
* configure: Rebuilt.
* src/powerpc/ffi.c (hidden): Define.
(ffi_prep_args_SYSV): Renamed from
ffi_prep_args. Cast pointers to unsigned long to shut up warnings.
(NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64,
ASM_NEEDS_REGISTERS64): New.
(ffi_prep_args64): New function.
(ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI.
(ffi_call): Likewise.
(ffi_prep_closure): Likewise.
(flush_icache): Surround by #ifndef POWERPC64.
(ffi_dblfl): New union type.
(ffi_closure_helper_SYSV): Use it to avoid aliasing problems.
(ffi_closure_helper_LINUX64): New function.
* src/powerpc/ppc_closure.S: Surround whole file by #ifndef
__powerpc64__.
* src/powerpc/sysv.S: Likewise.
(ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV.
* src/powerpc/linux64.S: New file.
* src/powerpc/linux64_closure.S: New file.
* Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and
src/powerpc/linux64_closure.S.
(TARGET_SRC_POWERPC): Likewise.
* src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2,
closure_test_fn3): Fix result printing on big-endian 64-bit
machines.
(main): Print tst2_arg instead of uninitialized tst2_result.
* src/ffitest.c (main): Hide what closure pointer really points to
from the compiler.
From-SVN: r65781
* libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32
bit mode.
* libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV):
Receive closure pointer through parameter, read args using
__builtin_dwarf_cfa.
(FFI_INIT_TRAMPOLINE): Send closure reference through eax.
From-SVN: r64663
2002-11-10 Ranjit Mathew <rmathew@hotmail.com>
* include/ffi.h.in: Added FFI_STDCALL ffi_type
enumeration for X86_WIN32.
* src/x86/win32.S: Added ffi_call_STDCALL function
definition.
* src/x86/ffi.c (ffi_call/ffi_raw_call): Added
switch cases for recognising FFI_STDCALL and
calling ffi_call_STDCALL if target is X86_WIN32.
* src/ffitest.c (my_stdcall_strlen/stdcall_many):
stdcall versions of the "my_strlen" and "many"
test functions (for X86_WIN32).
Added test cases to test stdcall invocation using
these functions.
From-SVN: r59878