Re: [PATCH] usb: musb: mark PM functions as __maybe_unused

From: Geert Uytterhoeven
Date: Mon Nov 28 2016 - 06:09:59 EST


Hi Arnd,

On Mon, Nov 28, 2016 at 11:54 AM, Arnd Bergmann <arnd@xxxxxxxx> wrote:
> On Monday, November 28, 2016 11:51:37 AM CET Geert Uytterhoeven wrote:
>> On Tue, Nov 22, 2016 at 3:30 PM, Arnd Bergmann <arnd@xxxxxxxx> wrote:
>> > Building without CONFIG_PM causes a harmless warning:
>> >
>> > drivers/usb/musb/musb_core.c:2041:12: error: âmusb_run_resume_workâ defined but not used [-Werror=unused-function]
>> >
>> > Removing the #ifdef around the PM code and instead marking the suspend/resume
>> > functions as __maybe_unused will do the right thing without warning.
>> >
>> > Fixes: ea2f35c01d5e ("usb: musb: Fix sleeping function called from invalid context for hdrc glue")
>> > Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
>>
>> linux-m68k-allmodconfig$ bloat-o-meter drivers/usb/musb/musb_core.o{.orig,}
>> add/remove: 8/0 grow/shrink: 0/0 up/down: 2344/0 (2344)
>> function old new delta
>> musb_restore_context - 892 +892
>> musb_save_context - 690 +690
>> musb_run_resume_work - 190 +190
>> musb_resume - 182 +182
>> musb_runtime_resume - 148 +148
>> musb_suspend - 114 +114
>> musb_dev_pm_ops - 92 +92
>> musb_runtime_suspend - 36 +36
>> Total: Before=13091, After=15435, chg +17.91%
>
> Well, in allmodconfig, you have CONFIG_PM enabled, so this is not
> dead code but actually does what was intended (though possibly
> not written as efficiently as it could have been).

M68k does not have CONFIG_PM.

> In a configuration without CONFIG_PM, there should ideally be
> no added code.

Apparently not :-(

BTW, same result for allyesconfig.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds