| Message ID | 20100917125740.GC4049@virgil.arch.suse.de |
|---|---|
| State | New |
| Headers | show |
On Fri, 17 Sep 2010, Martin Jambor wrote: > Hi, > > this is a followup to my previous build_ref_for_offset patch. If > build_user_friendly_ref_for_offset does not succeed when building an > artificial SRA access, we use build_ref_for_model which produces an > expression we cannot use in warnings and so I introduced a flag in the > access to remember that fact. > > Bootstrapped and tested on x86_64-linux, ok for trunk? Ok. Thanks, Richard. > Thanks, > > Martin > > > 2010-09-15 Martin Jambor <mjambor@suse.cz> > > * tree-sra.c (struct access): New field grp_no_warning. > (create_access_replacement): Set TREE_NO_WARNING according to > grp_no_warning. > (create_artificial_child_access): Use build_ref_for_model and set > grp_no_warning if build_user_friendly_ref_for_offset fails. > (propagate_subaccesses_across_link): Likewise. > > Index: mine/gcc/tree-sra.c > =================================================================== > --- mine.orig/gcc/tree-sra.c > +++ mine/gcc/tree-sra.c > @@ -216,6 +216,9 @@ struct access > cannot be called from within FOR_EACH_REFERENCED_VAR. */ > unsigned grp_to_be_replaced : 1; > > + /* Should TREE_NO_WARNING of a replacement be set? */ > + unsigned grp_no_warning : 1; > + > /* Is it possible that the group refers to data which might be (directly or > otherwise) modified? */ > unsigned grp_maybe_modified : 1; > @@ -1714,7 +1717,10 @@ create_access_replacement (struct access > } > SET_DECL_DEBUG_EXPR (repl, debug_expr); > DECL_DEBUG_EXPR_IS_FROM (repl) = 1; > - TREE_NO_WARNING (repl) = TREE_NO_WARNING (access->base); > + if (access->grp_no_warning) > + TREE_NO_WARNING (repl) = 1; > + else > + TREE_NO_WARNING (repl) = TREE_NO_WARNING (access->base); > } > else > TREE_NO_WARNING (repl) = 1; > @@ -1970,12 +1976,17 @@ create_artificial_child_access (struct a > tree expr = parent->base; > > gcc_assert (!model->grp_unscalarizable_region); > - if (!build_user_friendly_ref_for_offset (&expr, TREE_TYPE (expr), new_offset, > - model->type)) > - return NULL; > > access = (struct access *) pool_alloc (access_pool); > memset (access, 0, sizeof (struct access)); > + if (!build_user_friendly_ref_for_offset (&expr, TREE_TYPE (expr), new_offset, > + model->type)) > + { > + access->grp_no_warning = true; > + expr = build_ref_for_model (EXPR_LOCATION (parent->base), parent->base, > + new_offset, model, NULL, false); > + } > + > access->base = parent->base; > access->expr = expr; > access->offset = new_offset; > @@ -2016,11 +2027,16 @@ propagate_subaccesses_across_link (struc > { > tree t = lacc->base; > > + lacc->type = racc->type; > if (build_user_friendly_ref_for_offset (&t, TREE_TYPE (t), lacc->offset, > racc->type)) > + lacc->expr = t; > + else > { > - lacc->expr = t; > - lacc->type = racc->type; > + lacc->expr = build_ref_for_model (EXPR_LOCATION (lacc->base), > + lacc->base, lacc->offset, > + racc, NULL, false); > + lacc->grp_no_warning = true; > } > return false; > } > >
On Fri, Sep 17, 2010 at 5:57 AM, Martin Jambor <mjambor@suse.cz> wrote: > Hi, > > this is a followup to my previous build_ref_for_offset patch. If > build_user_friendly_ref_for_offset does not succeed when building an > artificial SRA access, we use build_ref_for_model which produces an > expression we cannot use in warnings and so I introduced a flag in the > access to remember that fact. > > Bootstrapped and tested on x86_64-linux, ok for trunk? > > Thanks, > > Martin > > > 2010-09-15 Martin Jambor <mjambor@suse.cz> > > * tree-sra.c (struct access): New field grp_no_warning. > (create_access_replacement): Set TREE_NO_WARNING according to > grp_no_warning. > (create_artificial_child_access): Use build_ref_for_model and set > grp_no_warning if build_user_friendly_ref_for_offset fails. > (propagate_subaccesses_across_link): Likewise. > This caused: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50444
Index: mine/gcc/tree-sra.c =================================================================== --- mine.orig/gcc/tree-sra.c +++ mine/gcc/tree-sra.c @@ -216,6 +216,9 @@ struct access cannot be called from within FOR_EACH_REFERENCED_VAR. */ unsigned grp_to_be_replaced : 1; + /* Should TREE_NO_WARNING of a replacement be set? */ + unsigned grp_no_warning : 1; + /* Is it possible that the group refers to data which might be (directly or otherwise) modified? */ unsigned grp_maybe_modified : 1; @@ -1714,7 +1717,10 @@ create_access_replacement (struct access } SET_DECL_DEBUG_EXPR (repl, debug_expr); DECL_DEBUG_EXPR_IS_FROM (repl) = 1; - TREE_NO_WARNING (repl) = TREE_NO_WARNING (access->base); + if (access->grp_no_warning) + TREE_NO_WARNING (repl) = 1; + else + TREE_NO_WARNING (repl) = TREE_NO_WARNING (access->base); } else TREE_NO_WARNING (repl) = 1; @@ -1970,12 +1976,17 @@ create_artificial_child_access (struct a tree expr = parent->base; gcc_assert (!model->grp_unscalarizable_region); - if (!build_user_friendly_ref_for_offset (&expr, TREE_TYPE (expr), new_offset, - model->type)) - return NULL; access = (struct access *) pool_alloc (access_pool); memset (access, 0, sizeof (struct access)); + if (!build_user_friendly_ref_for_offset (&expr, TREE_TYPE (expr), new_offset, + model->type)) + { + access->grp_no_warning = true; + expr = build_ref_for_model (EXPR_LOCATION (parent->base), parent->base, + new_offset, model, NULL, false); + } + access->base = parent->base; access->expr = expr; access->offset = new_offset; @@ -2016,11 +2027,16 @@ propagate_subaccesses_across_link (struc { tree t = lacc->base; + lacc->type = racc->type; if (build_user_friendly_ref_for_offset (&t, TREE_TYPE (t), lacc->offset, racc->type)) + lacc->expr = t; + else { - lacc->expr = t; - lacc->type = racc->type; + lacc->expr = build_ref_for_model (EXPR_LOCATION (lacc->base), + lacc->base, lacc->offset, + racc, NULL, false); + lacc->grp_no_warning = true; } return false; }