> Hi!
>
> In the following testcase, we predict baz to have cold
> entry regardless of the user supplied attribute (as it call
> unconditionally a cold function), but still issue
> a -Wsuggest-attribute=cold warning despite it having that attribute
> already.
>
> The following patch avoids that.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
>
> 2023-03-25 Jakub Jelinek <ja...@redhat.com>
>
> PR ipa/105685
> * predict.cc (compute_function_frequency): Don't call
> warn_function_cold if function already has cold attribute.
>
> * c-c++-common/cold-2.c: New test.
>
> --- gcc/predict.cc.jj 2023-01-02 09:32:38.273055726 +0100
> +++ gcc/predict.cc 2023-03-24 16:54:13.658606215 +0100
> @@ -4033,7 +4033,9 @@ compute_function_frequency (void)
> }
>
> node->frequency = NODE_FREQUENCY_UNLIKELY_EXECUTED;
> - warn_function_cold (current_function_decl);
> + if (lookup_attribute ("cold", DECL_ATTRIBUTES (current_function_decl))
> + == NULL)
> + warn_function_cold (current_function_decl);
OK, tanks!
In general we probably want to walk aliases and suggest warning on
aliases attached to the function, but we get this wrong with other
attributes too, so I will add it to TODo for next stage1.
Honza
> if (ENTRY_BLOCK_PTR_FOR_FN (cfun)->count.ipa() == profile_count::zero ())
> return;
> FOR_EACH_BB_FN (bb, cfun)
> --- gcc/testsuite/c-c++-common/cold-2.c.jj 2023-03-24 16:56:07.344000973
> +0100
> +++ gcc/testsuite/c-c++-common/cold-2.c 2023-03-24 16:55:58.985119001
> +0100
> @@ -0,0 +1,19 @@
> +/* PR ipa/105685 */
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -Wsuggest-attribute=cold" } */
> +
> +extern void foo (char *, char const *, int);
> +
> +__attribute__((cold)) char *
> +bar (int x)
> +{
> + static char b[42];
> + foo (b, "foo", x);
> + return b;
> +}
> +
> +__attribute__((cold)) char *
> +baz (int x) /* { dg-bogus "function might be candidate for
> attribute 'cold'" } */
> +{
> + return bar (x);
> +}
>
> Jakub
>
