* breakpoint.h (bp_location): Add related_address member.
* inferior.h (get_return_value): Take a pointer to struct value
instead of struct type for the function requested.
* value.h (using_struct_return): Likewise.
* gdbarch.sh (return_value): Take a pointer to struct value
instead of struct type for the function requested.
* breakpoint.c (set_breakpoint_location_function): Initialize
related_address for bp_gnu_ifunc_resolver breakpoints.
* elfread.c (elf_gnu_ifunc_resolver_return_stop): Pass the
requested function's address to gdbarch_return_value.
* eval.c (evaluate_subexp_standard): Pass the requested
function's address to using_struct_return.
* infcall.c (call_function_by_hand): Pass the requested
function's address to using_struct_return and
gdbarch_return_value.
* infcmd.c (get_return_value): Take a pointer to struct value
instead of struct type for the function requested.
(print_return_value): Update accordingly.
(finish_command_continuation): Likewise.
* stack.c (return_command): Pass the requested function's
address to using_struct_return and gdbarch_return_value.
* value.c (using_struct_return): Take a pointer to struct value
instead of struct type for the function requested. Pass the
requested function's address to gdbarch_return_value.
* python/py-finishbreakpoint.c (finish_breakpoint_object):
New function_value member, replacing function_type.
(bpfinishpy_dealloc): Update accordingly.
(bpfinishpy_pre_stop_hook): Likewise.
(bpfinishpy_init): Likewise. Record the requested function's
address.
* mips-tdep.c (mips_fval_reg): New enum.
(mips_o32_push_dummy_call): For MIPS16 FP doubles do not swap
words put in GP registers.
(mips_o64_push_dummy_call): Update a comment.
(mips_o32_return_value): Take a pointer to struct value instead
of struct type for the function requested and use it to check if
using the MIPS16 calling convention. Return the designated
general purpose registers for floating-point values returned in
MIPS16 mode.
(mips_o64_return_value): Likewise.
* ppc-tdep.h (ppc_sysv_abi_return_value): Update prototype.
(ppc_sysv_abi_broken_return_value): Likewise.
(ppc64_sysv_abi_return_value): Likewise.
* alpha-tdep.c (alpha_return_value): Take a pointer to struct
value instead of struct type for the function requested.
* amd64-tdep.c (amd64_return_value): Likewise.
* amd64-windows-tdep.c (amd64_windows_return_value): Likewise.
* arm-tdep.c (arm_return_value): Likewise.
* avr-tdep.c (avr_return_value): Likewise.
* bfin-tdep.c (bfin_return_value): Likewise.
* cris-tdep.c (cris_return_value): Likewise.
* frv-tdep.c (frv_return_value): Likewise.
* h8300-tdep.c (h8300_return_value): Likewise.
(h8300h_return_value): Likewise.
* hppa-tdep.c (hppa32_return_value): Likewise.
(hppa64_return_value): Likewise.
* i386-tdep.c (i386_return_value): Likewise.
* ia64-tdep.c (ia64_return_value): Likewise.
* iq2000-tdep.c (iq2000_return_value): Likewise.
* lm32-tdep.c (lm32_return_value): Likewise.
* m32c-tdep.c (m32c_return_value): Likewise.
* m32r-tdep.c (m32r_return_value): Likewise.
* m68hc11-tdep.c (m68hc11_return_value): Likewise.
* m68k-tdep.c (m68k_return_value): Likewise.
(m68k_svr4_return_value): Likewise.
* m88k-tdep.c (m88k_return_value): Likewise.
* mep-tdep.c (mep_return_value): Likewise.
* microblaze-tdep.c (microblaze_return_value): Likewise.
* mn10300-tdep.c (mn10300_return_value): Likewise.
* moxie-tdep.c (moxie_return_value): Likewise.
* mt-tdep.c (mt_return_value): Likewise.
* ppc-linux-tdep.c (ppc_linux_return_value): Likewise.
* ppc-sysv-tdep.c (ppc_sysv_abi_return_value): Likewise.
(ppc_sysv_abi_broken_return_value): Likewise.
(ppc64_sysv_abi_return_value): Likewise.
* ppcnbsd-tdep.c (ppcnbsd_return_value): Likewise.
* rl78-tdep.c (rl78_return_value): Likewise.
* rs6000-aix-tdep.c (rs6000_return_value): Likewise.
* rx-tdep.c (rx_return_value): Likewise.
* s390-tdep.c (s390_return_value): Likewise.
* score-tdep.c (score_return_value): Likewise.
* sh-tdep.c (sh_return_value_nofpu): Likewise.
(sh_return_value_fpu): Likewise.
* sh64-tdep.c (sh64_return_value): Likewise.
* sparc-tdep.c (sparc32_return_value): Likewise.
* sparc64-tdep.c (sparc64_return_value): Likewise.
* spu-tdep.c (spu_return_value): Likewise.
* tic6x-tdep.c (tic6x_return_value): Likewise.
* v850-tdep.c (v850_return_value): Likewise.
* vax-tdep.c (vax_return_value): Likewise.
* xstormy16-tdep.c (xstormy16_return_value): Likewise.
* xtensa-tdep.c (xtensa_return_value): Likewise.
* gdbarch.c: Regenerate.
* gdbarch.h: Regenerate.
gdb/testsuite/
* gdb.base/return-nodebug.exp: Also test float and double types.
(ppc_sysv_abi_push_dummy_call): Use it.
(do_ppc_sysv_return_value): Likewise.
(ppc64_sysv_abi_push_dummy_call): Likewise.
(ppc64_sysv_abi_return_value): Likewise.
Our testsuite noticed a crash when trying to call a function which
requires GDB to allocate memory in the inferior. Typically, this
happens when one of the parameters is a string. For instance, our
testcase tries:
(gdb) call debug.trace (me, "You")
[1] 32737 segmentation fault /path/to/gdb
What happens is that GDB sees the string, and thus tries to allocate
memory for it in the inferior:
> /* Allocate NBYTES of space in the inferior using the inferior's
> malloc and return a value that is a pointer to the allocated
> space. */
>
> struct value *
> value_allocate_space_in_inferior (int len)
> {
> struct objfile *objf;
> struct value *val = find_function_in_inferior ("malloc", &objf);
And find_function_in_inferior first searches the symtab in case
we have debug info. But, in our case (bareboard powerpc), we don't,
so it gets "malloc"'s address from the minimal symbols, and builds
a value whose type is a TYPE_CODE_PTR, not a TYPE_CODE_FUNC.
As a result, when we later try to make the call to malloc, we end up
inside the powerpc tdep code that has:
> do_ppc_sysv_return_value (struct gdbarch *gdbarch, struct type *func_type,
[...]
> if (func_type
> && TYPE_CALLING_CONVENTION (func_type) == DW_CC_GDB_IBM_OpenCL)
The problem is that func_type is not a TYPE_CODE_FUNC, and thus
the type-specific kind is not TYPE_SPECIFIC_FUNC, and so we do
TYPE_CALLING_CONVENTION is an invalid access.
Interestingly, the other call to TYPE_CALLING_CONVENTION is correctly
preceded by a check of the type's TYPE_CODE (making sure that it is
TYPE_CODE_FUNC).
gdb/ChangeLog:
* ppc-sysv-tdep.c (do_ppc_sysv_return_value): Do not check
FUNC_TYPE's calling convention if FUNC_TYPE is not a function.
* ppc-sysv-tdep.c (ppc64_sysv_abi_push_dummy_call): Add support
for the "generic" vector ABI used with GCC 4.3 and later.
(ppc64_sysv_abi_return_value): Likewise.
gdb/testsuite:
* gdb.arch/altivec-abi.exp: Skip "generic" tests on 64-bit when
using a GCC 4.1 or 4.2 compiler. Add an additional test variant
"generic ABI, auto".
(altivec_abi_tests): Accept vectors returned by reference.
* ppc-sysv-tdep.c (ppc64_sysv_abi_push_dummy_call): Implement
correct ABI for AltiVec vector arguments.
testsuite/ChangeLog:
* gdb.arch/altivec-abi.c (vec_func): Make use of intv_on_stack_f
when computing result.
* gdb.arch/altivec-abi.exp: Update expected results.
platforms that use function descriptors. Prefer to use function
pointer types instead of function types.
* linespec.c (decode_objc): Support function descriptors. Fully
initialize SAL result.
* objc-lang.c (find_methods): Support function descriptors.
Do not require function symbol to point to text section.
* ppc-sysv-tdep.c (ppc64_sysv_abi_push_dummy_call): When calling
via a function pointer, use the descriptor it points to.
decimal floating-point values in GPRs for soft-float.
(do_ppc_sysv_return_value): Handle returning decimal
floating-point values in GPRs for soft-float.
TYPE_CODE_BOOL and TYPE_CODE_CHAR the same as TYPE_CODE_INT.
Handle TYPE_CODE_REF the same as TYPE_CODE_PTR.
Handle TYPE_CODE_METHOD the same as TYPE_CODE_FUNC.
Allow typedefs when checking for function pointer arguments.
Right-align small structs passed on the stack.
(ppc64_sysv_abi_return_value): Handle TYPE_CODE_BOOL and
TYPE_CODE_CHAR the same as TYPE_CODE_INT.
Handle TYPE_CODE_REF the same as TYPE_CODE_PTR.
Thiago Jung Bauermann <bauerman@br.ibm.com>
* ppc-sysv-tdep.c (ppc_sysv_abi_push_dummy_call): Pass floats that
don't fit into registerson the stack the way GCC does.
TYPE_CODE_DECFLOAT arguments.
(ppc64_sysv_abi_push_dummy_call) Likewise.
(get_decimal_float_return_value): New function.
(do_ppc_sysv_return_value): Add support for TYPE_CODE_DECFLOAT return
values by calling get_decimal_float_return_value.
(ppc64_sysv_abi_return_value): Likewise.
Daniel Jacobowitz <dan@codesourcery.com>
* ppc-sysv-tdep.c (ppc64_sysv_abi_push_dummy_call): Right-align
struct values smaller than one doubleword; left-align those
larger. Pass structs containing a single floating-point value in
registers.
soft float and vector ABIs. Support the generic vector ABI for
AltiVec types.
(do_ppc_sysv_return_value): Likewise. Correct argument types and
casts.
(ppc64_sysv_abi_push_dummy_call): Assert that floating point is
supported.
* ppc-tdep.h (enum powerpc_vector_abi): New.
(struct gdbarch_tdep): Add soft_float and vector_abi.
* rs6000-tdep.c (setpowerpccmdlist, showpowerpccmdlist)
(powerpc_soft_float_global, powerpc_vector_strings)
(powerpc_vector_abi_global, powerpc_vector_abi_string): New.
(rs6000_gdbarch_init): Check for soft-float and vector ABI markings.
(set_powerpc_command, show_powerpc_command, powerpc_set_soft_float)
(powerpc_set_vector_abi): New.
(_initialize_rs6000_tdep): Register "set powerpc" and "show powerpc"
commands.
* Makefile.in (elf_ppc_h): New.
(rs6000-tdep.o): Update.
* gdb.texinfo (PowerPC): Document "set powerpc vector-abi" and "set
powerpc soft-float".
* gdb.arch/altivec-abi.exp: Run multiple times for GCC on GNU/Linux.
Test "set powerpc vector-abi". Skip auto-detection tests for old
toolchains.
instead of calling read_register.
* avr-tdep.c (avr_push_dummy_call): Write to REGCACHE instead of
calling write_register.
* hppa-tdep.c (hppa32_push_dummy_call): Write to REGCACHE instead of
calling write_register.
* ia64-tdep.c (find_func_descr): Add REGCACHE parameter. Use it
instead of calling read_register.
(ia64_push_dummy_call): Update call to find_func_descr. Use REGCACHE
instead of calling read_register and write_register.
* m32r-tdep.c (m32r_linux_supply_gregset): Use REGCACHE parameter
instead of current_regcache.
* mn10300-tdep.c (mn10300_push_dummy_call): Write to REGCACHE instead
of calling write_register.
* mn10300-linux-tdep.c (am33_supply_fpregset_method): Use REGCACHE
parameter instead of current_regcache.
* mips-tdep.c (mips2_fp_compat): Add FRAME parameter. Use it
instead of calling read_register.
(mips_read_fp_register_double, mips_print_fp_register): Update calls.
(mips_eabi_push_dummy_call): Use REGCACHE instead of write_register.
(mips_n32n64_push_dummy_call): Likewise.
(mips_o32_push_dummy_call): Likewise.
(mips_o64_push_dummy_call): Likewise.
* ppc-sysv-tdep.c (ppc_sysv_abi_push_dummy_call): Use REGCACHE
parameter instead of current_regcache.
* xtensa-tdep.c (xtensa_register_write_masked): Add REGCACHE parameter.
Use it instead of read_register and write_register.
(xtensa_register_read_masked): Likewise.
(xtensa_pseudo_register_read): Update call.
(xtensa_pseudo_register_write): Likewise.
(xtensa_frame_cache): Use register values unwound from NEXT_FRAME
instead of calling read_register.
(xtensa_push_dummy_call): Update comment.
* ppc-sysv-tdep.c (ppc64_sysv_abi_push_dummy_call)
(convert_code_addr_to_desc_addr): Convert any function code
address to the corresponding function's descriptor.
(ppc64_sysv_abi_return_value): have TYPE_CODE_ENUM and
TYPE_CODE_INT use the same code paths as TYPE_CODE_INT. When
writing, convert any function code address to the corresponding
descriptor.
* Makefile.in (ppc-sysv-tdep.o): Update dependencies.
* ppc-sysv-tdep.c: Include "objfiles.h".
(ppc64_sysv_abi_push_dummy_call): Provide the
the FN's objfile when looking for the descriptor.
