> On Mon, May 02, 2016 at 07:01:38PM +0200, Jan Hubicka wrote: > > This patch impleemnts the suggested check for presence of FP parameters. > > We can play with special casing the moves incrementally. > > This patch has broken the build: > > /home/marek/src/gcc/gcc/ipa-inline.c: In function ‘bool > can_inline_edge_p(cgraph_edge*, bool, bool, bool)’: > /home/marek/src/gcc/gcc/ipa-inline.c:341:55: error: ‘CIF_THUNK’ was not > declared in this scope > e->inline_failed = e->caller->thunk.thunk_p ? CIF_THUNK : > CIF_MISMATCHED_ARGUMENTS;
Sorry, I managed to miss independent changes that was in my tree. I will revert the evaluate_conditions_for_known_args and add CIF_THUNK shortly. Honza > ^ > Makefile:1085: recipe for target 'ipa-inline.o' failed > make: *** [ipa-inline.o] Error 1 > make: *** Waiting for unfinished jobs.... > /home/marek/src/gcc/gcc/ipa-inline-analysis.c: In function ‘clause_t > evaluate_conditions_for_known_args(cgraph_node*, bool, vec<tree_node*>, > vec<ipa_agg_jump_function*>)’: > /home/marek/src/gcc/gcc/ipa-inline-analysis.c:854:27: error: invalid > conversion > from ‘tree_node*’ to ‘long int’ [-fpermissive] > c->offset, c->by_ref); > ^ > /home/marek/src/gcc/gcc/ipa-inline-analysis.c:854:27: error: too many > arguments > to function ‘tree_node* ipa_find_agg_cst_for_param(ipa_agg_jump_function*, > long > int, bool)’ > In file included from /home/marek/src/gcc/gcc/ipa-inline-analysis.c:90:0: > /home/marek/src/gcc/gcc/ipa-prop.h:639:6: note: declared here > tree ipa_find_agg_cst_for_param (struct ipa_agg_jump_function *, > HOST_WIDE_INT, > ^ > Makefile:1085: recipe for target 'ipa-inline-analysis.o' failed > make: *** [ipa-inline-analysis.o] Error 1 > > Marek
