On April 5, 2016 5:03:54 PM GMT+02:00, Jeff Law <l...@redhat.com> wrote: >On 04/05/2016 08:59 AM, Cristina Georgiana Opriceana wrote: >> On Mon, Mar 28, 2016 at 10:55 PM, Richard Biener >> <richard.guent...@gmail.com> wrote: >>> On March 28, 2016 7:23:26 PM GMT+02:00, Cristina Georgiana Opriceana ><cristina.oprice...@gmail.com> wrote: >>>> Hello, >>>> >>>> In order to compute all the statements where a variable is used, is >it >>>> enough to rely on the SSA analysis? I tried to do something like >this: >>>> >>>> FOR_EACH_LOCAL_DECL (cfun, i, var) { >>>> for (unsigned int i = 0; i < num_ssa_names; i++) { >>>> if (ssa_name(i) && SSA_NAME_VAR (ssa_name(i)) == var) { >>>> tree ssa_node = ssa_name(i); >>>> FOR_EACH_IMM_USE_FAST (use_op, imm_iter, ssa_node) >{ >>>> gimple * use_stmt = USE_STMT(use_op); >>>> ... >>>> } >>> >>> No, you'll miss variables not in SSA form. >>> >> >> Could you give me more details on this or maybe some reading >material? >> As far as I understand, all the VAR_DECL, PARAM_DECL and RESULT_DECL >> will be in SSA form. >Only non-addressable _DECL nodes will be in SSA form. > >ie, arrays, or objects which have had their address taken will not be >in >SSA form.
All aggregates will not be in SSA form. Richard. >jeff