Move floating-point code to new function in i387-tdep.c.
* i387-tdep.c (i387_register_to_value, i387_value_to_register):
New functions containing code moved here from i386-tdep.c.
* i387-tdep.h: Add opaque declaration for `struct type'.
(i387_register_to_value, i387_value_to_register): New prototypes.
* x86-64-tdep.c (x86_64_convert_register_p): New function.
(x86_64_init_abi): Set convert_register_p, register_to_value and
value_to_register here.
Mark Kettenis <kettenis@gnu.org>
* gdbarch.sh (CONVERT_REGISTER_P): Add "type" parameter.
(REGISTER_TO_VALUE, VALUE_TO_REGISTER): Replace raw buffer
parameter with "frame".
* gdbarch.h, gdbarch.c: Re-generate.
* frame.h (put_frame_register): Declare.
* frame.c (put_frame_register): New function.
* arch-utils.c (legacy_convert_register_p): Add "type" parameter.
(legacy_register_to_value): Rewrite, use "frame" to get the
register value.
(legacy_value_to_register): Rewrite, use "frame" to find the
register's location before storing.
* arch-utils.h (legacy_convert_register_p): Update.
(legacy_register_to_value, legacy_value_to_register): Update.
* findvar.c (value_from_register): Rewrite, eliminate use of
REGISTER_CONVERT_TO_TYPE, pass "type" to CONVERT_REGISTER_P, pass
"frame" to REGISTER_TO_VALUE.
* valops.c (value_assign): Move the CONVERT_REGISTER code to the
lval_reg_frame_relative + lval_register branch of the switch. Do
not use REGISTER_CONVERT_FROM_TYPE. Use put_frame_register.
* i386-tdep.c (I386_EBX_REGNUM, I386_ECX_REGNUM, I386_ESI_REGNUM,
I386_EDI_REGNUM): New defines.
(i386_next_regnum, i386_convert_register_p,
i386_register_to_value, i386_value_to_register): New functions.
(i386_register_convertible, i386_register_convert_to_virtual,
i386_convert_to_raw): Remove functions.
(i386_gdbarch_init): Set convert_register_p, register_to_value and
value_to_register instead of register_convertible,
register_convert_to_virtual and register_convert_to_raw.
* mips-tdep.c (mips_convert_register_p): New function.
(mips_value_to_register): Replace mips_register_convert_from_type.
(mips_register_to_value): Replace mips_register_convert_to_type.
(mips_gdbarch_init): Set conver_register_p, value_to_register and
register_to_value.
* alpha-tdep.c (alpha_convert_register_p): Update.
(alpha_value_to_register): Update, store the register.
(alpha_register_to_value): Update, fetch the register.
* gdb.base/store.exp: Test longest and doublest. Test all
parameters. Weaken return statement match.
* gdb.base/store.c: Add longest and doublest - aka long long and
long double functions. Put all parameters into local register
variables. Use negative values.
(avr_gdbarch_init): Remove avr_call_dummy_words variable.
Don't set deprecated_call_dummy_words.
Remove commented out set_gdbarch_believe_pcc_promotion() call.
Don't set remote_translate_xfer_address.
(avr_io_reg_read_command): Remove commented out debug printf.
Wrap a long line.
* gdb.base/fileio.c: Include <errno.h>, and <sys/wait.h>. Gag
-Wformat errors. Add lost line. Use WEXITSTATUS to get system
exit status.
* gdb.base/fileio.exp: Disable target when nointerrupts and
noinferiorio, instead of limiting it to remote. Use remote_exec
instead of system.
* infcall.c (call_function_by_hand): When UNWIND_DUMMY_ID is
available, do not use the FP register, and always save the TOS.
* dummy-frame.c (dummy_frame_this_id): Do not assert
SAVE_DUMMY_FRAME_TOS.
* i386-tdep.c (i386_save_dummy_frame_tos): Delete function.
(i386_gdbarch_init): Do not set save_dummy_frame_tos.
(i386_push_dummy_call): Add 8 to the returned SP.
* frame.c (legacy_frame_p): Do not require SAVE_DUMMY_FRAME_TOS.
* d10v-tdep.c (d10v_unwind_dummy_id): Use d10v_unwind_sp.
(d10v_gdbarch_init): Do not set save_dummy_frame_tos.
* x86-64-tdep.c (x86_64_save_dummy_frame_tos): Delete function.
(x86_64_push_dummy_call): Return "sp + 16".
(x86_64_init_abi): Do not set save_dummy_frame_tos.
* alpha-tdep.c (alpha_gdbarch_init): Do not set
save_dummy_frame_tos.
(avr_pointer_to_address): Shift code addrs left 1 bit.
(avr_convert_from_func_ptr_addr): Delete function since operation is
better handled by avr_address_to_pointer and avr_pointer_to_address.
(avr_gdbarch_init): Don't set convert_from_func_ptr_add method.
(FPU_DEFAULT, case TE_NetBSD): Default to FPU_ARCH_VFP for ELF,
FPU_ARCH_FPA for AOUT.
(md_begin): Don't try to guess the floating point architecture from
the CPU if the OS ABI (Linux, NetBSD) mandates a particular form.
indicate registers used for return values.
(struct frame_extra_info): Drop args_pointer and locals_pointer.
(h8300_examine_prologue): Remove initializing dropped frame_extra_info
members.
(h8300_init_extra_frame_info): Ditto.
(h8300_frame_locals_address): Removed.
(h8300_frame_args_address): Removed.
(h8300_extract_return_value): Use new regcache structure. Only care
for 16 bit CPUs.
(h8300h_extract_return_value): Same function for 32 bit CPUs.
(h8300_store_return_value): Use new regcache structure. Only care
for 16 bit CPUs.
(h8300h_store_return_value): Same function for 32 bit CPUs.
(h8300_store_struct_return): Removed.
(h8300_extract_struct_value_address): Use new regcache structure.
(h8300h_extract_struct_value_address): Removed.
(h8300_push_dummy_code): New function.
(h8300_gdbarch_init): Slightly rearranged to stress deprecated calls.
Remove call_dummy_words. Call set_gdbarch_extract_return_value and
set_gdbarch_store_return_value architecture dependent.
Call set_gdbarch_push_dummy_code and
set_gdbarch_extract_struct_value_address.
Remove calls to set_gdbarch_frame_args_address,
set_gdbarch_frame_locals_address,
set_gdbarch_deprecated_store_struct_return,
set_gdbarch_deprecated_extract_return_value,
set_gdbarch_deprecated_extract_struct_value_address,
set_gdbarch_deprecated_call_dummy_words and
set_gdbarch_deprecated_sizeof_call_dummy_words.
(h8300h_register_byte): Remove.
(h8300_register_virtual_type): Remove. Substitute by...
(h8300_register_type): New function.
(h8300_extract_struct_value_address): Drop usage of h8300_register_byte.
(h8300h_extract_struct_value_address): Ditto.
(h8300_gdbarch_init): Drop calls to
set_gdbarch_deprecated_register_byte and
set_gdbarch_deprecated_register_virtual_type.
Add call to set_gdbarch_register_type.
* ppc-linux-tdep.c (insn_d, insn_ds, insn_xfx, read_insn, struct
insn_pattern, insns_match_pattern, d_field, ds_field): New
functions, macros, and types for working with PPC instructions.
(ppc64_standard_linkage, PPC64_STANDARD_LINKAGE_LEN,
ppc64_in_solib_call_trampoline, ppc64_standard_linkage_target,
ppc64_skip_trampoline_code): New functions, variables, and macros
for recognizing and skipping linkage functions.
(ppc_linux_init_abi): Use ppc64_in_solib_call_trampoline and
ppc64_skip_trampoline_code for the 64-bit PowerPC Linux ABI.
* config/powerpc/tm-linux.h (SKIP_TRAMPOLINE_CODE): Remove the
other #definition of this.
(ppc_linux_skip_trampoline_code): Remove declaration.
* ppc-linux-tdep.c (ppc_linux_skip_trampoline_code): Make this
static.
(ppc_linux_init_abi): Register it as the skip_trampoline_code
method for GDBARCH.
powerpc64-*-linux, selecting powerpc/ppc64-linux.mh and
powerpc/linux.mt.
* config/powerpc/ppc64-linux.mh, config/powerpc/nm-ppc64-linux.mh:
New files.
* wince.c: Include "mips-tdep.h".
* mips-tdep.h (mips_next_pc): Declare.
* mcore-tdep.c: Make more local functions static.
* Makefile.in (wince.o): Update dependencies.
