On Tue, Aug 09, 2016 at 05:21:26PM +0200, Marek Polacek wrote:
> Bootstrapped/regtested on x86_64-linux and ppc64le-redhat-linux.
>
> 2016-08-09 Marek Polacek <pola...@redhat.com>
>
> PR c/7652
> gcc/c-family/
> * c-ada-spec.c (dump_generic_ada_node): Add return.
> gcc/
> * cselib.c (cselib_expand_value_rtx_1): Add return.
> * gengtype.c (dbgprint_count_type_at): Likewise.
> * hsa-gen.c (gen_hsa_insn_for_internal_fn_call): Likewise.
> * reg-stack.c (get_true_reg): Restructure to avoid fallthrough warning.
>
> --- gcc/gcc/cselib.c
> +++ gcc/gcc/cselib.c
> @@ -1618,6 +1618,7 @@ cselib_expand_value_rtx_1 (rtx orig, struct
> expand_value_data *evd,
> else
> return orig;
> }
> + return orig;
> }
>
> CASE_CONST_ANY:
> --- gcc/gcc/gengtype.c
> +++ gcc/gcc/gengtype.c
> @@ -175,6 +175,7 @@ dbgprint_count_type_at (const char *fil, int lin, const
> char *msg, type_p t)
> {
> case TYPE_UNDEFINED:
> nb_undefined++;
> + break;
> case TYPE_SCALAR:
> nb_scalar++;
> break;
These two are ok.
> --- gcc/gcc/reg-stack.c
> +++ gcc/gcc/reg-stack.c
> @@ -423,23 +423,25 @@ get_true_reg (rtx *pat)
> GET_MODE (subreg));
> return pat;
> }
> + pat = &XEXP (*pat, 0);
> + break;
> }
While this one is similar to the cselib.c case, I wonder if it wouldn't be
nicer to drop the {}s, reindent the case SUBREG: and just /* FALLTHRU */
at the end. Uninitialized decls can be declared anywhere in the switch
in C++ or in C99+. I don't feel strongly about it, so if you don't want to
do that, this hunk is ok too.
> case FLOAT:
> case FIX:
> case FLOAT_EXTEND:
> - pat = & XEXP (*pat, 0);
> + pat = &XEXP (*pat, 0);
> break;
>
> case UNSPEC:
> if (XINT (*pat, 1) == UNSPEC_TRUNC_NOOP
> || XINT (*pat, 1) == UNSPEC_FILD_ATOMIC)
> - pat = & XVECEXP (*pat, 0, 0);
> + pat = &XVECEXP (*pat, 0, 0);
> return pat;
>
> case FLOAT_TRUNCATE:
> if (!flag_unsafe_math_optimizations)
> return pat;
> - pat = & XEXP (*pat, 0);
> + pat = &XEXP (*pat, 0);
> break;
>
> default:
The above 3 are obviously ok.
Jakub
