I've fixed the misprint and vertical spaces.
I'll ask to commit the patch when x86 bootstrap and make check finished.
Thanks,
Evgeny
Updated ChangeLog:
2015-10-30 Evgeny Stupachenko <evstu...@gmail.com>
gcc/
* Makefile.in (OBJS): Add multiple_target.o.
* attrib.c (make_attribute): Moved from config/i386/i386.c
* config/i386/i386.c (make_attribute): Deleted.
* multiple_target.c (create_dispatcher_calls): New.
(get_attr_len): Ditto.
(get_attr_str): Ditto.
(separate_attrs): Ditto.
(is_valid_asm_symbol): Ditto.
(create_new_asm_name): Ditto.
(create_target_clone): Ditto.
(expand_target_clones): Ditto.
(ipa_target_clone): Ditto.
(ipa_dispatcher_calls): Ditto.
* passes.def (pass_target_clone): Two new ipa passes.
* tree-pass.h (make_pass_target_clone): Ditto.
gcc/c-family
* c-common.c (handle_target_clones_attribute): New.
* (c_common_attribute_table): Add handle_target_clones_attribute.
* (handle_always_inline_attribute): Add check on target_clones
attribute.
* (handle_target_attribute): Ditto.
gcc/testsuite
* gcc.dg/mvc1.c: New test for multiple targets cloning.
* gcc.dg/mvc2.c: Ditto.
* gcc.dg/mvc3.c: Ditto.
* gcc.dg/mvc4.c: Ditto.
* gcc.dg/mvc5.c: Ditto.
* gcc.dg/mvc6.c: Ditto.
* gcc.dg/mvc7.c: Ditto.
* g++.dg/ext/mvc1.C: Ditto.
* g++.dg/ext/mvc2.C: Ditto.
* g++.dg/ext/mvc3.C: Ditto.
* g++.dg/ext/mvc4.C: Ditto.
gcc/doc
* doc/extend.texi (target_clones): New attribute description.
On Fri, Oct 30, 2015 at 8:27 AM, Jeff Law <l...@redhat.com> wrote:
> On 10/29/2015 12:13 PM, Evgeny Stupachenko wrote:
>>
>> On Thu, Oct 29, 2015 at 8:02 PM, Jan Hubicka<hubi...@ucw.cz> wrote:
>>>>
>>>> >>Yes. This is not necessary. However that way we'll have the following
>>>> >>code in dispatcher:
>>>> >> cmpl $6, __cpu_model+4(%rip)
>>>> >> sete %al
>>>> >> movzbl %al, %eax
>>>> >> testl %eax, %eax
>>>> >> jle .L16
>>>> >> movl $foo.target_clone.1, %eax
>>>> >>I think it is very hard to read and debug such...
>>>> >>
>>>> >>While now we have:
>>>> >>
>>>> >> cmpl $6, __cpu_model+4(%rip)
>>>> >> sete %al
>>>> >> movzbl %al, %eax
>>>> >> testl %eax, %eax
>>>> >> jle .L16
>>>> >> movl $foo.arch_slm, %eax
>>>> >>
>>>> >>and it is clear that we are jumping to SLM code here.
>>>> >>I'd like to keep target in names.
>>>
>>> >
>>> >I am not against more informative names, but why you don't pass the info
>>> > here:
>>> >
>>> >+create_target_clone (cgraph_node *node, bool definition)
>>> >+{
>>> >+ cgraph_node *new_node;
>>> >+ if (definition)
>>> >+ {
>>> >+ new_node = node->create_version_clone_with_body (vNULL, NULL,
>>> >+ NULL, false,
>>> >+ NULL, NULL,
>>> >+ "target_clone");
>>> >+ new_node->force_output = true;
>>> >+ }
>>> >+ else
>>> >+ {
>>> >+ tree new_decl = copy_node (node->decl);
>>> >+ new_node = cgraph_node::get_create (new_decl);
>>> >+ }
>>> >+ return new_node;
>>> >+}
>>> >
>>> >passing "arch_slm" instead of target_clone will get you the name you
>>> > want
>>> >(plus the extra index that may be needed anyway to disambiguate).
>>> >
>>> >Note that in general those .suffixes should be machine parseable, so
>>> > cp-demangle.c
>>> >can expand them correctly. We may want to have some consistent grammar
>>> > for them here
>>> >and update cp-demangle.c to output nice info like "target clone for..."
>>
>> Ok. I've modified the patch correspondingly.
>
> You'll need updated ChangeLog entries. Don't forget to drop the omp-low
> spurious whitespace change.
>
> You should also fix the formatting nits Jan pointed out.
>
> With those changes, this patch is OK for the trunk. I'll run the header
> file reordering & cleanup tool after the patch is committed to the trunk.
>
> jeff
>
>
