[i386] Fix PR 57756

On Mon, Sep 23, 2013 at 4:09 AM, Richard Biener
<richard.guenther@gmail.com> wrote:
> On Sat, Sep 21, 2013 at 4:09 AM, Sriraman Tallam <tmsriram@google.com> wrote:
>> Forgot to add the test case. Patch updated.
>>
>> Sri
>>
>> On Fri, Sep 20, 2013 at 7:06 PM, Sriraman Tallam <tmsriram@google.com> wrote:
>>> On Wed, Aug 14, 2013 at 3:38 AM, Richard Biener
>>> <richard.guenther@gmail.com> wrote:
>>>> On Wed, Aug 14, 2013 at 2:02 AM, Sriraman Tallam <tmsriram@google.com>
>>>> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>>     I have attached a patch to fix PR57756.  Description:  The
>>>>> following program,
>>>>>
>>>>> __attribute__((always_inline,target("sse4.2")))
>>>>> __inline int callee ()
>>>>> {
>>>>>   return 0;
>>>>> }
>>>>>
>>>>> __attribute__((target("sse")))
>>>>> __inline int caller ()
>>>>> {
>>>>>   return callee();
>>>>> }
>>>>>
>>>>> does not generate an error and callee is inlined into caller. This is
>>>>> because callee has a higher target ISA.  Interchanging the position of
>>>>> caller and callee will generate the correct error. Also, removing the
>>>>> target attribute from caller will generate the error.
>>>>>
>>>>> The reason for the bug is that when the caller's target options are
>>>>> processed, global_options contain the ISA flags of the callee
>>>>> (previously processed) and doing this in i386-common.c, where opts is
>>>>> global_options:
>>>>>
>>>>> opts->x_ix86_isa_flags |= OPTION_MASK_ISA_SSE_SET;
>>>>>
>>>>> is not changing global options.  The fix is to reset global_options to
>>>>> the default each time a new target option needs to be processed.
>>>>
>>>>
>>>> Shouldn't  ix86_valid_target_attribute_tree be refactored to work on a
>>>> selected
>>>> opt structure rather than global_options?
>>>
>>> I have attached a patch that does the refactoring of
>>> ix86_valid_target_attribute_tree. In that process I had to refactor
>>> ix86_valid_target_attribute_inner_p and ix86_option_override_internal.
>>> This made it a large patch because global_options structure is touched
>>> by these functions in multiple places and through vvarious aliases
>>> like ix86_arch_string for instance. In effect, this frees these
>>> functions from depending on global_options and the code in
>>> ix86_alid_target_attribute_p shows how the original bug can be fixed
>>> by just using a different gcc_options structure.
>
> This looks nice.  I suppose you grepped for effected targets and rs6000
> was the only one besides i386.
>
> This is ok given target maintainers do not object within 24h and the test
> successfully bootstrapped and passed regression tests.

I changed this patch  quite a bit after I realized I missed many other
places where global_options field was accessed. The two specific
changes are:

* Particularly, ix86_function_specific_save and
ix86_function_specific_restore had to be changed to copy to a specific
gcc_options structure than to global_options.
* Function ix86_option_override_internal accesses global_options via
macros, like TARGET_64BIT etc. This had to be changed. So, I created
new macros to accept a parameter and named them like TARGET_64BIT_P
and replaced all the accesses in i386.c

I also marked as TODO a copy in function ix86_function_specific_save.
Here, the cl_target_option structure has a ix86_target_flags_explicit
field whereas the global_options structure does not have one.
Previously, this used to get saved to the global_options target_flags
field but this did not make sense to me. I have commented this line
out. Please let me know if a new field needs to be added.

This patch passes bootstrap and all tests. Please take another look.

Sri

>
> Thanks,
> Richard.
>
>>> Please review.
>>>
>>> Thanks
>>> Sri
>>>
>>>
>>>
>>>> Richard.
>>>>
>>>>>
>>>>> Patch ok?
>>>>>
>>>>> Thanks
>>>>> Sri
>>>>
>>>>
* testsuite/gcc.target/i386/pr57756.c: New test.
	* optc-save-gen.awk: Add extra parameter to the save and restore
	target calls.
	* opth-gen.awk: Generate new TARGET_* macros  to accept a parameter.
	* tree.c (build_optimization_node): New parameter.  Add extra parameter
	to call to cl_optimization_save.
	(build_target_option_node): New parameter. Add extra parameter
	to call to cl_target_option_save.
	* tree.h (build_optimization_node): New parameter.
	(build_target_option_node): New parameter.
	* c-family/c-common.c (handle_optimize_attribute): Fix calls to
	build_optimization_node and build_target_option_node.
	* c-family/c-pragma.c (handle_pragma_optimize): Ditto.
	(handle_pragma_push_options): Ditto.
	* toplev.c (process_options): Ditto.
	* opts.c (init_options_struct): Check for opts_set non-null.
	* target.def (target_option.save): New parameter.
	(target_option.restore): New parameter.
	* tm.texi: Generate.
	* config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
	(ix86_pragma_target_parse): Ditto.
	* config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
	parameters.
	* config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
	to  build_optimization_node and build_target_option_node.
	(rs6000_valid_attribute_p): Ditto.
	(rs6000_pragma_target_parse): Ditto.
	* config/i386/i386.h:
	TARGET_64BIT_P: New Macro
	TARGET_MMX_P: New Macro.
	TARGET_3DNOW_P: New Macro.
	TARGET_3DNOW_A_P: New Macro.
	TARGET_SSE_P: New Macro.
	TARGET_SSE2_P: New Macro.
	TARGET_SSE3_P: New Macro.
	TARGET_SSSE3_P: New Macro.
	TARGET_SSE4_1_P: New Macro.
	TARGET_SSE4_2_P: New Macro.
	TARGET_AVX_P: New Macro.
	TARGET_AVX2_P: New Macro.
	TARGET_AVX512F_P: New Macro.
	TARGET_AVX512PF_P: New Macro.
	TARGET_AVX512ER_P: New Macro.
	TARGET_AVX512CD_P: New Macro.
	TARGET_FMA_P: New Macro.
	TARGET_SSE4A_P: New Macro.
	TARGET_FMA4_P: New Macro.
	TARGET_XOP_P: New Macro.
	TARGET_LWP_P: New Macro.
	TARGET_ABM_P: New Macro.
	TARGET_BMI_P: New Macro.
	TARGET_BMI2_P: New Macro.
	TARGET_LZCNT_P: New Macro.
	TARGET_TBM_P: New Macro.
	TARGET_POPCNT_P: New Macro.
	TARGET_SAHF_P: New Macro.
	TARGET_MOVBE_P: New Macro.
	TARGET_CRC32_P: New Macro.
	TARGET_AES_P: New Macro.
	TARGET_PCLMUL_P: New Macro.
	TARGET_CMPXCHG16B_P: New Macro.
	TARGET_FSGSBASE_P: New Macro.
	TARGET_RDRND_P: New Macro.
	TARGET_F16C_P: New Macro.
	TARGET_RTM_P: New Macro.
	TARGET_HLE_P: New Macro.
	TARGET_RDSEED_P: New Macro.
	TARGET_PRFCHW_P: New Macro.
	TARGET_ADX_P: New Macro.
	TARGET_FXSR_P: New Macro.
	TARGET_XSAVE_P: New Macro.
	TARGET_XSAVEOPT_P: New Macro.
	TARGET_LP64_P: New Macro.
	TARGET_X32_P: New Macro.
	TARGET_FPMATH_DEFAULT_P: New Macro.
	TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
	* config/i386/i386.c (ix86_option_override_internal): New parameters.
	opts and opts_set.
	Change ix86_tune_string to access opts->x_ix86_tune_string.
	Change ix86_isa_flags to access opts->x_ix86_isa_flags.
	Change ix86_arch_string to access opts->x_ix86_arch_string.
	Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
	Change ix86_pmode to access opts->x_ix86_pmode.
	Change ix86_abi to access opts->x_ix86_abi.
	Change ix86_cmodel to access opts->x_ix86_cmodel.
	Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
	Change ix86_isa_flags_explicit to access
	opts->x_ix86_isa_flags_explicit.
	Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
	Change ix86_regparm to access opts->x_ix86_regparm.
	Change ix86_branch_cost to access opts->x_ix86_branch_cost.
	Change ix86_preferred_stack_boundary_arg to access
	opts->x_ix86_preferred_stack_boundary_arg.
	Change ix86_force_align_arg_pointer to access
	opts->x_ix86_force_align_arg_pointer.
	Change ix86_incoming_stack_boundar_arg to access
	opts->x_ix86_incoming_stack_boundar_arg.
	Change ix86_fpmath to access opts->x_ix86_fpmath.
	Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
	Change ix86_recip_name to access opts->x_ix86_recip_name.
	Change ix86_stack_protector_guard to access
	opts->x_ix86_stack_protector_guard.
	Change ix86_tune_memcpy_strategy to access
	opts->x_ix86_tune_memcpy_strategy.
	Change ix86_tune_memset_strategy to access
	opts->x_ix86_tune_memset_strategy.
	Change global_options to access opts.
	Change global_options_set to access opts_set.
        Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
        Change TARGET_MMX to TARGET_MMX_P (opts->...)
        Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...)
        Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...)
        Change TARGET_SSE to TARGET_SSE_P (opts->...)
        Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
        Change TARGET_SSE3 to TARGET_SSE3_P (opts->...)
        Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...)
        Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...)
        Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...)
        Change TARGET_AVX to TARGET_AVX_P (opts->...)
        Change TARGET_AVX2 to TARGET_AVX2_P (opts->...)
        Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...)
        Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...)
        Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...)
        Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...)
        Change TARGET_FMA to TARGET_FMA_P (opts->...)
        Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...)
        Change TARGET_FMA4 to TARGET_FMA4_P (opts->...)
        Change TARGET_XOP to TARGET_XOP_P (opts->...)
        Change TARGET_LWP to TARGET_LWP_P (opts->...)
        Change TARGET_ABM to TARGET_ABM_P (opts->...)
        Change TARGET_BMI to TARGET_BMI_P (opts->...)
        Change TARGET_BMI2 to TARGET_BMI2_P (opts->...)
        Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...)
        Change TARGET_TBM to TARGET_TBM_P (opts->...)
        Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...)
        Change TARGET_SAHF to TARGET_SAHF_P (opts->...)
        Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...)
        Change TARGET_CRC32 to TARGET_CRC32_P (opts->...)
        Change TARGET_AES to TARGET_AES_P (opts->...)
        Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...)
        Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...)
        Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...)
        Change TARGET_RDRND to TARGET_RDRND_P (opts->...)
        Change TARGET_F16C to TARGET_F16C_P (opts->...)
        Change TARGET_RTM to TARGET_RTM_P (opts->...)
        Change TARGET_HLE to TARGET_HLE_P (opts->...)
        Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...)
        Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...)
        Change TARGET_ADX to TARGET_ADX_P (opts->...)
        Change TARGET_FXSR to TARGET_FXSR_P (opts->...)
        Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...)
        Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...)
        Change TARGET_LP64 to TARGET_LP64_P (opts->...)
        Change TARGET_X32 to TARGET_X32_P (opts->...)
        Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...)
        Change TARGET_FLOAT_RETURNS_IN_80387 to
	 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...)
	(ix86_function_specific_save): New parameter. Use opts-> fields
	to replace global fields.
	(ix86_function_specific_restore): Ditto.
	(ix86_valid_target_attribute_inner_p): New parameters.
	Fix recursive call.
	Fix call to ix86_handle_option and set_option.
	(ix86_valid_target_attribute_tree): New parameters.
	Change global_options to access opts.
	Change global_options_set to access opts_set.
	Fix call to ix86_valid_target_attribute_inner_p.
	Change ix86_tune_string to access opts->x_ix86_tune_string.
	Change ix86_arch_string to access opts->x_ix86_arch_string.
	Change ix86_fpmath to access opts->x_ix86_fpmath
	Fix call to ix86_option_override_internal.
	Fix call to ix86_add_new_builtins.
	Fix calls to build_optimization_node and build_target_option_node.
	(ix86_valid_target_attribute_p): Remove access to global_options.
	Use new gcc_options structure func_options.
	Fix call to ix86_valid_target_attribute_tree.
	Fix call to  build_optimization_node.
	(get_builtin_code_for_version):	Fix call to
	ix86_valid_target_attribute_tree.

[i386] Fix PR 57756

Commit Message

Patch