Re: [announce] new tree: "fix all build warnings, on all configs"

[Ingo Molnar <mingo@elte.hu>]

* Andi Kleen <andi@firstfloor.org> wrote:

> >   *	acpi_pm_disable_gpes - Disable the GPEs.
> >   */
> > -static int acpi_pm_disable_gpes(void)
> > +static inline int acpi_pm_disable_gpes(void)
> 
> Just to satisfy my curiosity, what compiler warning does marking 
> functions inline fix?

the commit log below explains the situation. The warning exposed a maze 
of #ifdefs in drivers/acpi/sleep/main.c. It's not the warning we need to 
"fix" but that maze, obviously.

	Ingo

-------------------------------------------->
>From 6ddae344a73fcff60c840dd4e429bf55562b41f3 Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@elte.hu>
Date: Fri, 17 Oct 2008 15:44:22 +0200
Subject: [PATCH]  #ifdef complications in drivers/acpi/sleep/main.c

this warning:

  drivers/acpi/sleep/main.c:67: warning: ‘acpi_pm_disable_gpes’ defined but not used
  drivers/acpi/sleep/main.c:92: warning: ‘acpi_pm_prepare’ defined but not used
  drivers/acpi/sleep/main.c:107: warning: ‘acpi_pm_finish’ defined but not used
  drivers/acpi/sleep/main.c:128: warning: ‘acpi_pm_end’ defined but not used

Shows that this code has an identity crisis due to a maze of #ifdefs, in
the PM_SLEEP && !SUSPEND && !HIBERNATION case for example.

Instead of complicating the code with even more #ifdefs, one option
would be to convert these rather trivial wrappers to inline functions
(implemented in this commit). Maybe there's a better solution as well -
such as to reduce the many config options we have in this area?

No size difference with SUSPEND && HIBERNATION turned back on:

drivers/acpi/sleep/main.o:

   text	   data	    bss	    dec	    hex	filename
   2717	    976	     33	   3726	    e8e	main.o.before
   2717	    976	     33	   3726	    e8e	main.o.after

md5:
   44220906eff0ea6e1a3266b35dd82ac2  main.o.before.asm
   44220906eff0ea6e1a3266b35dd82ac2  main.o.after.asm
---
 drivers/acpi/sleep/main.c |   10 +++++-----
 1 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/acpi/sleep/main.c b/drivers/acpi/sleep/main.c
index d13194a..2276d75 100644
--- a/drivers/acpi/sleep/main.c
+++ b/drivers/acpi/sleep/main.c
@@ -63,7 +63,7 @@ void __init acpi_old_suspend_ordering(void)
 /**
  *	acpi_pm_disable_gpes - Disable the GPEs.
  */
-static int acpi_pm_disable_gpes(void)
+static inline int acpi_pm_disable_gpes(void)
 {
 	acpi_hw_disable_all_gpes();
 	return 0;
@@ -75,7 +75,7 @@ static int acpi_pm_disable_gpes(void)
  *	If necessary, set the firmware waking vector and do arch-specific
  *	nastiness to get the wakeup code to the waking vector.
  */
-static int __acpi_pm_prepare(void)
+static inline int __acpi_pm_prepare(void)
 {
 	int error = acpi_sleep_prepare(acpi_target_sleep_state);
 
@@ -88,7 +88,7 @@ static int __acpi_pm_prepare(void)
  *	acpi_pm_prepare - Prepare the platform to enter the target sleep
  *		state and disable the GPEs.
  */
-static int acpi_pm_prepare(void)
+static inline int acpi_pm_prepare(void)
 {
 	int error = __acpi_pm_prepare();
 
@@ -103,7 +103,7 @@ static int acpi_pm_prepare(void)
  *	This is called after we wake back up (or if entering the sleep state
  *	failed).
  */
-static void acpi_pm_finish(void)
+static inline void acpi_pm_finish(void)
 {
 	u32 acpi_state = acpi_target_sleep_state;
 
@@ -124,7 +124,7 @@ static void acpi_pm_finish(void)
 /**
  *	acpi_pm_end - Finish up suspend sequence.
  */
-static void acpi_pm_end(void)
+static inline void acpi_pm_end(void)
 {
 	/*
 	 * This is necessary in case acpi_pm_finish() is not called during a