Make Sparc a Multi-Arch target. Discard PARAMS macro (require ANSI).
* sparc-tdep.c: include arch-utils.h.
(SPARC_HAS_FPU, FP_REGISTER_BYTES, FP_MAX_REG NUM, SPARC_INTREG_SIZE,
DUMMY_REG_SAVE_OFFSET): provide multi-arch-compatible definitions.
(GDB_TARGET_IS_SPARC64): make into a runtime test.
(struct frame_extra_info): Define, use instead of the macro.
(Many places): Use alloca instead of statically allocated buffers
that depend on a multi-arch variable such as MAX_REGISTER_RAW_SIZE.
(sparc_extract_struct_value_address): Accept a pointer arg instead
of an array sized by REGISTER_BYTES.
(examine_prologue): Accept a pointer to an array of CORE_ADDR,
instead of the defunct struct frame_saved_regs. Recognize new
Sparc64 store instructions as part of the prologue. Ignore the
destination of a frame store when parsing the prologue (so long
as it's on the stack).
(sparc_push_dummy_frame): Fix incorrect buffer offset for PSTATE.
(sparc_frame_find_saved_regs): Accept a ptr to an array of CORE_ADDR
instead of the defunct struct frame_saved_regs.
(supply_gregset): Discard unnecessary 'zerobuf': just send NULL to
supply_register. Provide 4-byte offset to compensate for diff
between size of the prgreg_t elements on a 64-bit host and size
of the registers for a 32-bit target. Fill all inaccessible regs
with zero so they won't keep being requested again and again.
(fill_gregset): Handle 32/64 size difference between registers
and prgreg_t. Handle as many new 64-bit regs as possible.
(supply_fpregset, fill_fpregset): Attempt to handle 64-bit world.
(sparc_push_arguments): Rename to sparc32_push_arguments.
Copy arguments into registers as well as onto stack, so that the
CALL_DUMMY (code pushed onto the target stack) is not required.
(sparc_extract_return_value): Rename to sparc32_extract_return_value.
(sparc_store_return_value): Use memset instead of bzero.
Use write_register_gen instead of write_register_bytes.
(sparclet_store_return_value): New function.
(_initialize_sparc_tdep): Call register_gdbarch_init to activate
the gdbarch multi-architecture system.
(sp64_push_arguments): Rename to sparc64_push_arguments.
Extend to store arguments in general registers as well as on stack.
(sparc64_extract_return_value): Rename to sp64_extract_return_value.
Use as a private function, to be called by the new external function
sparc64_extract_return_value.
(sparclet_extract_return_value): New function.
(sparc32_stack_align, sparc64_stack_align, sparc32_register_name,
sparc64_register_name, sparc_print_extra_frame_info,
sparclite_register_name, sparclet_register_name,
sparc_push_return_address, sparc64_use_struct_convention,
sparc32_store_struct_return, sparc64_store_struct_return,
sparc32_register_virtual_type, sparc64_register_virtual_type,
sparc32_register_size, sparc64_register_size,
sparc32_register_byte, sparc64_register_byte,
sparc_gdbarch_skip_prologue, sparc_convert_to_virtual,
sparc_convert_to_raw, sparc_frame_init_saved_regs,
sparc_frame_address, sparc_gdbarch_fix_call_dummy,
sparc_coerce_float_to_double, sparc_call_dummy_address,
sparc_y_regnum, sparc_reg_struct_has_addr, sparc_intreg_size,
sparc_return_value_on_stack): New functions supporting multi-arch.
(sparc_gdbarch_init): New function; initialize multi-arch.
(struct gdbarch_tdep): Define, use for private multi-arch data.
* config/sparc/tm-sparc.h: Move definitions around, enclose with
#ifdef GDB_MULTI_ARCH tests, provide some multi-arch alternate
definitions. Add enums for register names, to help debugging gdb.
This header file must work for non-multi-arch and for multi-arch.
* config/sparc/tm-sp64.h: Add GDB_MULTI_ARCH configuration. Also add
AT_ENTRY_POINT definitions for CALL_DUMMY, for non-multi-arch case.
Define GDB_MULTI_ARCH.
* config/sparc/tm-sparclet.h: Add GDB_MULTI_ARCH configuration.
Do not define GDB_MULTI_ARCH (bfd does not correctly identify target).
* config/sparc/tm-sparclite.h: Ditto.
* config/sparc/tm-sun4sol2.h: Define GDB_MULTI_ARCH.
* sparclet-rom.c (sparclet_regnames): Initialize explicitly, to
avoid using deprecated REGISTER_NAMES macro.
* Makefile.in: Let sparc-tdep.c depend on arch-utils.h.
* gdbarch.sh (IEEE_FLOAT): New entry.
* gdbarch.c, gdbarch.h: Regenerated.
* valprint.c (IEEE_FLOAT): Provide a default #definition for this.
(print_floating): Use IEEE_FLOAT as if it were an expression; use
the code specific to IEEE-format numbers whenever the value of
IEEE_FLOAT is non-zero.
* config/a29k/tm-a29k.h, config/alpha/tm-alpha.h,
config/arc/tm-arc.h, config/arm/tm-arm.h, config/fr30/tm-fr30.h,
config/h8300/tm-h8300.h, config/i386/tm-i386.h,
config/i960/tm-i960.h, config/m88k/tm-m88k.h,
config/mips/tm-mips.h, config/pa/tm-hppa.h,
config/sparc/tm-sparc.h, config/delta/tm-delta.h,
config/frv/tm-frv.h (IEEE_FLOAT): For all ports that #define
IEEE_FLOAT, make sure they give it the value (1).
* valops.c (COERCE_FLOAT_TO_DOUBLE): Rework definition to be
more function-like.
(default_coerce_float_to_double, standard_coerce_float_to_double):
New functions.
(value_arg_coerce): Adjust for new definition.
* value.h (default_coerce_float_to_double,
standard_coerce_float_to_double): New declarations for the above.
* gdbarch.sh (coerce_float_to_double): New entry, replacing macro.
* gdbarch.c, gdbarch.h: Regenerated.
* tm-alpha.h, tm-fr30.h, tm-m32r.h, tm-mips.h, tm-hppa.h,
tm-rs6000.h, tm-sh.h, tm-sparc.h (COERCE_FLOAT_TO_DOUBLE): Change
definitions.
* mips-tdep.c (mips_coerce_float_to_double): Supply our own custom
function here.
(mips_gdbarch_init): Install that as our coerce_float_to_double
function.
Add two pointers (saved_regs, extra_info) to struct frame_info.
Introduce new macro FRAME_INIT_SAVED_REGS which replaces
FRAME_FIND_SAVED_REGS.
Document.
Use in mn10300 and rs6000 targets. Fix side effects on ALPHA, MIPS,
Z8K and SPARC targets.
* sparc-tdep.c (sparc_extract_struct_value_address): Simplify to use
same method on both 32-bit and 64-bit machines.
* sparcl-tdep.c (sparclite_check_watch_resources): Simulator doesn't
support hardware breakpoints.
* config/sparc/tm-sparc.h (CALL_DUMMY): Improve comments.
set endianness based on machine type.
(_initialize_sparc_tdep): Initialize target_architecture_hook.
(sparc_print_register_hook): Print PSR and FPSR in fancy format
on 32-bit machines.
* config/sparc/tm-sparc.h (PRINT_REGISTER_HOOK): Redefine to
call sparc_print_register_hook instead of using inline code.
* config/sparc/tm-sp64.h (PRINT_REGISTER_HOOK): Remove.
%o0-%o5 as 64-bit values; compensate for stack bias.
(USE_STRUCT_CONVENTION): We only pass pointers to structs
if they're larger than 32 bytes.
(REG_STRUCT_HAS_ADDR): Ditto.
* sparc-tdep.c (sparc_init_extra_frame_info): Use read_sp()
instead of read_register. If the target is a sparc64 and the frame
pointer is odd, compensate for the stack bias.
(get_saved_register): Use read_sp().
(DUMMY_STACK_REG_BUF_SIZE): Use FP_REGISTER_BYTES.
(sparc_push_dummy_frame): Use read_sp()/write_sp(). On sparc64,
save the PC, NPC, CCR, FSR, FPRS, Y and ASI registers.
(sparc_frame_find_saved_regs): Use read_sp(). Read the PC, NPC,
CCR, FSR, FPRS, Y and ASI registers from the frame, if it's a
dummy frame.
(sparc_pop_frame): Use write_sp(). If the target is a sparc64 and
the FP is odd, compensate for stack bias.
(sparc_store_return_value): Right-justify the return value before
writing it to %o0.
(sparc_fix_call_dummy): Don't NOP out part of the call dummy on
sparc64.
(sparc64_read_sp, sparc64_read_fp, sparc64_write_sp,
sparc64_write_fp, sp64_push_arguments,
sparc64_extract_return_value): New functions to support the
sparc64 ABI.
* dwarfread.c (handle_producer): Set processing_gcc_compilation to
the right version number.
* dwarf2read.c (read_file_scope): Assume we're processing
GCC2 output.
make it work on the simulator.
(FIX_CALL_DUMMY): Convert to function call instead of inline code.
(sparc_fix_call_dummy): Declare.
* sparc-tdep.c (sparc_fix_call_dummy): New function, taken from
old FIX_CALL_DUMMY macro, with additional fixes for simulator.
(sparc_push_dummy_frame): Set registers differently on simulator
to prevent corrupted register window save areas.
(sparc_extract_return_value): New function, required to handle
machines without floating point.
(sparc_store_return_value): Ditto.
* config/sparc/tm-sparc.h (EXTRACT_RETURN_VALUE): Call
sparc_extract_return_value instead of using inline code.
(sparc_extract_return_value): Declare.
(STORE_RETURN_VALUE): Call sparc_store_return_value instead
of using inline code.
(sparc_store_return_value): Declare.
work correctly on little-endian hosts.
(sparc_push_arguments): New function.
(gdb_print_insn_sparc): New function.
(_initialize_sparc_tdep): Make gdb_print_insn_sparc the default
disassembler, so that SPARClite-specific instructions will
be recognized.
* sparcl-tdep.c (readchar): Print debugging information.
(debug_serial_write): New function, a replacement for SERIAL_WRITE
that prints debugging information.
* config/sparc/tm-sparc.h (PUSH_ARGUMENTS): Define.
(sparc_push_arguments): Declare.
Thu Jul 25 19:41:31 1996 Fred Fish <fnf@cygnus.com>
for details about these changes. I was going to include the complete
entry in the checkin message but it was apparently so large it caused
cvs to dump core.
tm_print_insn_info.
* gdbtk.c (gdb_disassemble): Setup di.mach from
tm_print_insn_info.mach, and set endian from TARGET_BYTE_ORDER.
* i386-tdep.c (set_assembly_language_command): set
tm_print_insn_info.mach to the appropriate value for 386 or 8086
disassembly.
* printcmd.c (print_insn): Move init of disassembler_info to
_initialize_printcmd. Set endian for disassembler here.
* sparc-tdep.c: Set tm_print_insn_info.mach as appropriate to
select sparc/sparclite.
* config/sparc/{tm-sparc.h tm-sparclite.h}: Get rid of
TM_PRINT_INSN. Set TM_PRINT_INSN_MACH to
bfd_mach_sparc/bfd_mach_sparc_sparclite.
config/sparc/tm-sparclite.h: Initialize tm_print_insn from
TM_PRINT_INSN, which comes from the tm file.
* Makefile.in (INTERNAL_LDFLAGS): Add in flags from configure.
* configure configure.in: Only make sol-thread.o for native.
Also, switch to dlopened libthread_db.so.1.
* sol-thread.c: Switch to using dlopen to get the thread_db
library.
* configure, configure.in: Change test for libthread_db to only
work for configs where build/host/target are the same.
* Makefile.in (gdbtk.o): Use X11_CFLAGS to provide alternate
locations (per-host) for X11 include files.
* config/pa/hppahpux.mh (XM_CLIBS): Add -L/usr/lib/X11R5 to force
the use of R5 libs. (X11_CFLAGS): Add this to indicate the locs
of the R5 include files.
end-sanitize-gdbtk
* monitor.c monitor.h remote-est.c rom68k-rom.c: Add start of
support for interrupting target.
* monitor.c (monitor_open): Send stop command before doing
anything else.
* (monitor_load_srec): Fix record size calculation to prevent end
of segment from getting trashed.
* rom68k-rom.c: Update to latest version of struct monitor_ops.
* config/sparc/tm-sparc.h (FIX_CALL_DUMMY): Fix byte-order
problems. Makes DOS hosted function calling work.
* sparclite/crt0.s: Define _start to make COFF happy.
* testsuite/config/rom68k.exp (gdb_target_rom68k): Use
$targetname, $serialport and $baud instead of hardwired variables.
* testsuite/gdb.base/{sigall.exp signals.exp}: Skip these if the
target doesn't support signals.
* sparc-tdep.c (sparc_init_extra_frame_info): Set it.
(examine_prologue, sparc_init_extra_frame_info): Use ->frame plus
->sp_offset to compute the address something is saved at, not
->bottom.
* sparc-tdep.c (get_saved_register): New function.
* tm-sparc.h: Define GET_SAVED_REGISTER; don't define
FRAME_FIND_SAVED_REGS, HAVE_REGISTER_WINDOWS or REGISTER_IN_WINDOW_P.
* stack.c (frame_info): Add comment about what to do if
FRAME_FIND_SAVED_REGS is not defined.
* sparc-tdep.c (sparc_init_extra_frame_info): Set ->frame field
here. Get it right for flat frames.
* sparc-tdep.c (sparc_frame_chain): Instead of returning
meaningful value for ->frame field, just return dummy value.
This change is needed because the old code didn't deal with mixed
flat and non-flat frames.
* sparc-tdep.c (sparc_pop_frame): Write SP_REGNUM from
frame->frame, don't go through saved regs for this.
* sparc-tdep.c: Move guts of skip_prologue to new function
examine_prologue. Check for flat prologue and set is_flat.
Provide the caller with the information about what is saved where
if desired.
(skip_prologue, sparc_frame_find_saved_regs): Call examine_prologue.
* sparc-tdep.c: Replace union sparc_insn_layout and anonymous
union in isannulled, which won't work on a little-endian host,
with X_* macros.
* sparc-tdep.c (sparc_frame_saved_pc): If addr == 0, the saved PC
is still in %o7.
* config/sparc/tm-sparc.h: Define INIT_FRAME_PC and
INIT_FRAME_PC_FIRST.
* blockframe.c (get_prev_frame_info): Modify comments regarding
INIT_FRAME_PC_FIRST and the sparc.
* sparc-tdep.c (single_step): Use 4 not sizeof (long) for size of
instruction.
for the structure's type. All callers changed.
* valops.c (call_function_by_hand): Check REG_STRUCT_HAS_ADDR
for each structure argument rather than assuming it's either
true or false for all structure arguments.
* config/pa/tm-hppa.h (REG_STRUCT_HAS_ADDR): Depend only
on the length structure passed, not the compiler used.
* config/sparc/tm-sparc.h (REG_STRUCT_HAS_ADDR): Accept additional
argument for the structure's type.
convert a LOC_ARG to a LOC_REF_ARG. Update code which combines
'p' and 'r' symbol descriptors into a single symbol to look for a
LOC_REF_ARG.
* README, config/sparc/tm-sparc.h: Update comments.
hppa-tdep.c (pa_print_fp_reg), infcmd.c (do_registers_info),
valops.c (value_assign): Use REGISTER_CONVERT_TO_* only if
REGISTER_CONVERTIBLE is defined, otherwise just copy the content.
Pass desired type to REGISTER_CONVERT_TO_*.
* config/m68k/tm-m68k.h, config/i960/tm-i960.h (REGISTER_CONVERT_*):
Pass length of desired type to store/extract_floating.
* config/i386/tm-arm.h, config/i386/tm-i386aix.h,
config/i386/tm-sun386.h, config/i386/tm-symmetry.h,
config/m88k/tm-m88k.h config/rs6000/tm-rs6000.h (REGISTER_CONVERT_*):
Use extract_floating and store_floating with length of desired type.
* config/m68k/tm-news.h (STORE,EXTRACT_RETURN_VALUE): Add type
parameter to REGISTER_CONVERT_*.
* config/a29k/tm-a29k.h, config/convex/tm-convex.h,
config/gould/tm-np1.h, config/gould/tm-pn.h, config/h8300/tm-h8300.h,
config/h8500/tm-h8500.h, config/i386/tm-i386v.h,
config/mips/tm-mips.h, config/ns32k/tm-merlin.h,
config/ns32k/tm-umax.h, config/pa/tm-hppa.h, config/pyr/tm-pyr.h,
config/sh/tm-sh.h, config/sparc/tm-sparc.h, config/tahoe/tm-tahoe.h,
config/vax/tm-vax.h, config/z8k/tm-z8k.h (REGISTER_CONVERTIBLE,
REGISTER_CONVERT_TO_RAW, REGISTER_CONVERT_TO_VIRTUAL): Remove
versions for which REGISTER_CONVERTIBLE is always false.
* z8k-tdep.c (register_convert_to_virtual, register_convert_to_raw):
Remove, no longer used.
* alpha-tdep.c (alpha_register_convert_to_raw,
alpha_register_convert_to_virtual): New routines to handle
the different raw formats in alpha floating point registers.
* config/alpha/tm-alpha.h (REGISTER_CONVERTIBLE,
REGISTER_CONVERT_TO_VIRTUAL, REGISTER_CONVERT_TO_RAW): Use them.
Some of the following is in #ifdef CALL_DUMMY_BREAKPOINT_OFFSET.
* breakpoint.h (enum bptype): Add bp_call_dummy.
(struct bpstat_what): Add call_dummy field.
* infrun.c (wait_for_inferior): Deal with it.
* breakpoint.c (bpstat_what): Deal with call dummy breakpoint.
* infcmd.c (run_stack_dummy): Set the call dummy breakpoint.
* config/sparc/tm-sparc.h: Define CALL_DUMMY_BREAKPOINT_OFFSET.
command. Now we use unwind info by default if we can find it.
* config/sparc: Move VARIABLES_INSIDE_BLOCK and SUN_FIXED_LBRAC_BUG
to tm-sparc.h so they are shared between Solaris and SunOS4.
* dbxread.c (process_one_symbol): Deal with SunOS4 acc N_STSYM and
N_GSYM functions.
