On Thu, Apr 12, 2018 at 3:50 PM, H.J. Lu <hjl.to...@gmail.com> wrote: > On Thu, Apr 12, 2018 at 6:39 AM, H.J. Lu <hjl.to...@gmail.com> wrote: >> On Thu, Apr 12, 2018 at 5:17 AM, Jan Hubicka <hubi...@ucw.cz> wrote: >>>> On Thu, Apr 12, 2018 at 1:29 PM, H.J. Lu <hongjiu...@intel.com> wrote: >>>> > Since IFUNC resolver is called indirectly, don't mark IFUNC resolver as >>>> > only called directly. >>>> > >>>> > OK for trunk? >>>> > >>>> > >>>> > H.J. >>>> > --- >>>> > gcc/ >>>> > >>>> > PR target/85345 >>>> > * cgraph.h: Include stringpool.h" and "attribs.h". >>>> > (cgraph_node::only_called_directly_or_aliased_p): Return false >>>> > for IFUNC resolver. >>>> > >>>> > gcc/testsuite/ >>>> > >>>> > PR target/85345 >>>> > * gcc.target/i386/pr85345.c: New test. >>>> > --- >>>> > gcc/cgraph.h | 5 +++- >>>> > gcc/testsuite/gcc.target/i386/pr85345.c | 44 >>>> > +++++++++++++++++++++++++++++++++ >>>> > 2 files changed, 48 insertions(+), 1 deletion(-) >>>> > create mode 100644 gcc/testsuite/gcc.target/i386/pr85345.c >>>> > >>>> > diff --git a/gcc/cgraph.h b/gcc/cgraph.h >>>> > index d1ef8408497..9e195824fcc 100644 >>>> > --- a/gcc/cgraph.h >>>> > +++ b/gcc/cgraph.h >>>> > @@ -24,6 +24,8 @@ along with GCC; see the file COPYING3. If not see >>>> > #include "profile-count.h" >>>> > #include "ipa-ref.h" >>>> > #include "plugin-api.h" >>>> > +#include "stringpool.h" >>>> > +#include "attribs.h" >>>> > >>>> > class ipa_opt_pass_d; >>>> > typedef ipa_opt_pass_d *ipa_opt_pass; >>>> > @@ -2894,7 +2896,8 @@ cgraph_node::only_called_directly_or_aliased_p >>>> > (void) >>>> > && !DECL_STATIC_CONSTRUCTOR (decl) >>>> > && !DECL_STATIC_DESTRUCTOR (decl) >>>> > && !used_from_object_file_p () >>>> > - && !externally_visible); >>>> > + && !externally_visible >>>> > + && !lookup_attribute ("ifunc", DECL_ATTRIBUTES (decl))); >>>> >>>> How's it handled for our own generated resolver functions? That is, >>>> isn't there sth cheaper than doing a lookup_attribute here? I see >>>> that make_dispatcher_decl nor ix86_get_function_versions_dispatcher >>>> adds the 'ifunc' attribute (though they are TREE_PUBLIC there). >>> >>> Is there any drawback of setting force_output flag? >>> Honza >> >> Setting force_output may prevent some optimizations. Can we add a bit >> for IFUNC resolver? >> > > Here is the patch to add ifunc_resolver to cgraph_node. Tested on x86-64 > and i686. Any comments? >
PING: https://gcc.gnu.org/ml/gcc-patches/2018-04/msg00647.html -- H.J.