Re: [3/3] 2.6.23-rc2: known regressions v2

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Monday, 13 August 2007 20:35, Sam Ravnborg wrote:
> On Mon, Aug 13, 2007 at 03:21:54PM +0200, Michal Piotrowski wrote:
> > > >
> > > > The patch changes the warning to a different warning, it's now
> > > > kernel/power/Kconfig:79:warning: 'select' used by config symbol 'PM_SLEEP_SMP' refers to undefined symbol 'HOTPLUG_CPU'
> 
> So we have something like this for ppc:
> 
> config PM_SLEEP_SMP
> 	bool "bla bla"

Plus a couple of "depends on ..." here, which cannot be true.

> 	select HOTPLUG_CPU
> 
> As it is not in my -linus tree it has been added by the referanced patch
> in the text I zapped.
> 
> 
> The warning is generated _independently_ of the actual configuration.
> All it says is that the symbol PM_SLEEP_SMP has a seelct statement
> for an unknown symbol - unknown for this architecture at least.
> 
> Grepping will reveal that HOTPLUG_CPU is not present for ppc and
> the easiet fix here seems to just add it like in the following:
> 
> diff --git a/arch/ppc/Kconfig b/arch/ppc/Kconfig
> index 6bdeeb7..e7bcd57 100644
> --- a/arch/ppc/Kconfig
> +++ b/arch/ppc/Kconfig
> @@ -65,6 +65,10 @@ config GENERIC_BUG
>         default y
>         depends on BUG
>  
> +config HOTPLUG_CPU
> +       bool
> +       default n
> +
>  source "init/Kconfig"
>  
>  menu "Processor"
> 
> [cut'n'paste so whitespace damaged...]
> This will silence the warning...
> 
> 
> Now as PM_SLEEP_SMP requires HOTPLUG_CPU there most be done
> something else to prevent this config symbol to be selected
> otherwise we would end up with a kernel that has HOTPLUG_CPU
> defined even it is not supported.

Well, the issue at hand is that PM_SLEEP_SMP will never be set on this arch
and the warning is printed nevertheless.

Greetings,
Rafael


-- 
"Premature optimization is the root of all evil." - Donald Knuth