https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116142
--- Comment #12 from rguenther at suse dot de <rguenther at suse dot de> --- On Tue, 6 Aug 2024, xry111 at gcc dot gnu.org wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116142 > > --- Comment #11 from Xi Ruoyao <xry111 at gcc dot gnu.org> --- > (In reply to rguent...@suse.de from comment #10) > > On Thu, 1 Aug 2024, xry111 at gcc dot gnu.org wrote: > > > > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116142 > > > > > > --- Comment #9 from Xi Ruoyao <xry111 at gcc dot gnu.org> --- > > > In the following call stack: > > > > > > #0 single_imm_use (var=0x7ffff564a3a8, use_p=0x7fffffffc9d8, > > > stmt=0x7fffffffc9e0) at ../../gcc/gcc/ssa-iterators.h:427 > > > #1 0x0000000001aabcde in vec_info::lookup_single_use (this=0x33a1b80, > > > lhs=0x7ffff564a3a8) at ../../gcc/gcc/tree-vectorizer.cc:581 > > > #2 0x0000000001a1a40a in supportable_widening_operation > > > (vinfo=0x33a1b80, > > > code=..., stmt_info=0x3482b50, vectype_out=0x7ffff595a1f8, > > > vectype_in=0x7ffff595a000, code1=0x7fffffffcb90, > > > code2=0x7fffffffcb94, > > > multi_step_cvt=0x7fffffffcba0, interm_types=0x7fffffffcc48) > > > at ../../gcc/gcc/tree-vect-stmts.cc:14196 > > > #3 0x00000000019f7c2d in vectorizable_conversion (vinfo=0x33a1b80, > > > stmt_info=0x3482b50, gsi=0x0, vec_stmt=0x0, slp_node=0x33b4cc8, > > > cost_vec=0x7fffffffd190) at ../../gcc/gcc/tree-vect-stmts.cc:5438 > > > #4 0x0000000001a16ecc in vect_analyze_stmt (vinfo=0x33a1b80, > > > stmt_info=0x3482b50, need_to_vectorize=0x7fffffffcf73, > > > node=0x33b4cc8, > > > node_instance=0x32fb6e0, cost_vec=0x7fffffffd190) > > > at ../../gcc/gcc/tree-vect-stmts.cc:13291 > > > > > > "ptr == ptr->next" is satisfied (the comment says it means "there aren't > > > any > > > uses whatsoever" but I don't really understand it: I knows almost nothing > > > about > > > tree vectorization) and single_imm_use directly returns false. > > > > So likely > > > > tree lhs = gimple_assign_lhs (stmt_info->stmt); > > stmt_vec_info use_stmt_info = loop_info->lookup_single_use > > (lhs); > > if (use_stmt_info > > > > needs to be > > > > tree lhs = gimple_assign_lhs (vect_orig_stmt (stmt_info)->stmt); > > stmt_vec_info use_stmt_info = loop_info->lookup_single_use > > (lhs); > > Works for me! Can you produce, test and post a patch?
