Re: [2.6.31.13] Build failure at arch/x86/kernel/apic/io_apic.c

Re: [2.6.31.13] Build failure at arch/x86/kernel/apic/io_apic.c

[Tetsuo Handa]

Commit d539e5576605d048e6aeb21cbe3a8e71dc5eea81 "x86: Fix SCI on IOAPIC != 0"
introduced "void setup_IO_APIC_irq_extra(u32 gsi)" which calls
mp_find_ioapic() and mp_find_ioapic_pin().

This commit does not compile if CONFIG_X86_IO_APIC=y and CONFIG_ACPI=n .

Regards.

Greg KH wrote:
> On Mon, Apr 12, 2010 at 11:30:43AM +0900, Tetsuo Handa wrote:
> > Hello.
> > 
> >   CC      arch/x86/kernel/apic/io_apic.o
> > arch/x86/kernel/apic/io_apic.c: In function `setup_IO_APIC_irq_extra':
> > arch/x86/kernel/apic/io_apic.c:1575: error: implicit declaration of function `mp_find_ioapic'
> > arch/x86/kernel/apic/io_apic.c:1579: error: implicit declaration of function `mp_find_ioapic_pin'
> > make[3]: *** [arch/x86/kernel/apic/io_apic.o] Error 1
> > make[2]: *** [arch/x86/kernel/apic] Error 2
> > make[1]: *** [arch/x86/kernel] Error 2
> > make: *** [arch/x86] Error 2
> > 
> > This is because
> > 
> > #ifdef CONFIG_ACPI
> > (...snipped...)
> > #ifdef CONFIG_X86_IO_APIC
> > extern int mp_find_ioapic(int gsi);
> > extern int mp_find_ioapic_pin(int ioapic, int gsi);
> > #endif
> > (...snipped...)
> > #endif
> > 
> > but CONFIG_X86_IO_APIC=y and CONFIG_ACPI=n in my config.
> > 2.6.31.12 was OK.
> 
> What patch caused this?
> 
> thanks,
> 
> greg k-h

Re: [2.6.31.13] Build failure at arch/x86/kernel/apic/io_apic.c

[Greg KH]

On Mon, Apr 12, 2010 at 01:10:59PM +0900, Tetsuo Handa wrote:
> Commit d539e5576605d048e6aeb21cbe3a8e71dc5eea81 "x86: Fix SCI on IOAPIC != 0"
> introduced "void setup_IO_APIC_irq_extra(u32 gsi)" which calls
> mp_find_ioapic() and mp_find_ioapic_pin().
> 
> This commit does not compile if CONFIG_X86_IO_APIC=y and CONFIG_ACPI=n .

Why would you want to build without ACPI on any modern system?  Anyway,
have a fix for this?

thanks,

greg k-h

Re: [2.6.31.13] Build failure at arch/x86/kernel/apic/io_apic.c

[Yinghai]

On 05/01/2010 11:07 AM, Greg KH wrote:
> On Mon, Apr 12, 2010 at 01:10:59PM +0900, Tetsuo Handa wrote:
>> Commit d539e5576605d048e6aeb21cbe3a8e71dc5eea81 "x86: Fix SCI on IOAPIC != 0"
>> introduced "void setup_IO_APIC_irq_extra(u32 gsi)" which calls
>> mp_find_ioapic() and mp_find_ioapic_pin().
>>
>> This commit does not compile if CONFIG_X86_IO_APIC=y and CONFIG_ACPI=n .
> 
> Why would you want to build without ACPI on any modern system?  Anyway,
> have a fix for this?
> 
maybe just put #ifdef CONFIG_ACPI around the callee etc?

YH

Re: [2.6.31.13] Build failure at arch/x86/kernel/apic/io_apic.c

[Greg KH]

On Mon, May 03, 2010 at 10:16:06AM -0700, Yinghai wrote:
> On 05/01/2010 11:07 AM, Greg KH wrote:
> > On Mon, Apr 12, 2010 at 01:10:59PM +0900, Tetsuo Handa wrote:
> >> Commit d539e5576605d048e6aeb21cbe3a8e71dc5eea81 "x86: Fix SCI on IOAPIC != 0"
> >> introduced "void setup_IO_APIC_irq_extra(u32 gsi)" which calls
> >> mp_find_ioapic() and mp_find_ioapic_pin().
> >>
> >> This commit does not compile if CONFIG_X86_IO_APIC=y and CONFIG_ACPI=n .
> > 
> > Why would you want to build without ACPI on any modern system?  Anyway,
> > have a fix for this?
> > 
> maybe just put #ifdef CONFIG_ACPI around the callee etc?

Ick.  How was it solved upstream?

thanks,

greg k-h

Re: [2.6.31.13] Build failure at arch/x86/kernel/apic/io_apic.c

[Yinghai]

On 05/03/2010 10:44 AM, Greg KH wrote:
> On Mon, May 03, 2010 at 10:16:06AM -0700, Yinghai wrote:
>> On 05/01/2010 11:07 AM, Greg KH wrote:
>>> On Mon, Apr 12, 2010 at 01:10:59PM +0900, Tetsuo Handa wrote:
>>>> Commit d539e5576605d048e6aeb21cbe3a8e71dc5eea81 "x86: Fix SCI on IOAPIC != 0"
>>>> introduced "void setup_IO_APIC_irq_extra(u32 gsi)" which calls
>>>> mp_find_ioapic() and mp_find_ioapic_pin().
>>>>
>>>> This commit does not compile if CONFIG_X86_IO_APIC=y and CONFIG_ACPI=n .
>>>
>>> Why would you want to build without ACPI on any modern system?  Anyway,
>>> have a fix for this?
>>>
>> maybe just put #ifdef CONFIG_ACPI around the callee etc?
> 
> Ick.  How was it solved upstream?
> 

those two functions are moved to arch/x86/kernel/apic/io_apic.c

maybe you can put that patch to 2.6.31.xx?

YH

commit 2a4ab640d3c28c2952967e5f63ea495555bf2a5f
Author: Feng Tang <feng.tang@intel.com>
Date:   Tue Jul 7 23:01:15 2009 -0400

    ACPI, x86: expose some IO-APIC routines when CONFIG_ACPI=n
    
    Some IO-APIC routines are ACPI specific now, but need to
    be exposed when CONFIG_ACPI=n for the benefit of SFI.
    
    Remove #ifdef ACPI around these routines:
    
    io_apic_get_unique_id(int ioapic, int apic_id);
    io_apic_get_version(int ioapic);
    io_apic_get_redir_entries(int ioapic);
    
    Move these routines from ACPI-specific boot.c to io_apic.c:
    
    uniq_ioapic_id(u8 id)
    mp_find_ioapic()
    mp_find_ioapic_pin()
    mp_register_ioapic()
    
    Also, since uniq_ioapic_id() is now no longer static,
    re-name it to io_apic_unique_id() for consistency
    with the other public io_apic routines.
    
    For simplicity, do not #ifdef the resulting code ACPI || SFI,
    thought that could be done in the future if it is important
    to optimize the !ACPI !SFI IO-APIC x86 kernel for size.
    
    Signed-off-by: Feng Tang <feng.tang@intel.com>
    Signed-off-by: Len Brown <len.brown@intel.com>
    Cc: x86@kernel.org

Re: [2.6.31.13] Build failure at arch/x86/kernel/apic/io_apic.c

[Greg KH]

On Mon, May 03, 2010 at 10:50:41AM -0700, Yinghai wrote:
> On 05/03/2010 10:44 AM, Greg KH wrote:
> > On Mon, May 03, 2010 at 10:16:06AM -0700, Yinghai wrote:
> >> On 05/01/2010 11:07 AM, Greg KH wrote:
> >>> On Mon, Apr 12, 2010 at 01:10:59PM +0900, Tetsuo Handa wrote:
> >>>> Commit d539e5576605d048e6aeb21cbe3a8e71dc5eea81 "x86: Fix SCI on IOAPIC != 0"
> >>>> introduced "void setup_IO_APIC_irq_extra(u32 gsi)" which calls
> >>>> mp_find_ioapic() and mp_find_ioapic_pin().
> >>>>
> >>>> This commit does not compile if CONFIG_X86_IO_APIC=y and CONFIG_ACPI=n .
> >>>
> >>> Why would you want to build without ACPI on any modern system?  Anyway,
> >>> have a fix for this?
> >>>
> >> maybe just put #ifdef CONFIG_ACPI around the callee etc?
> > 
> > Ick.  How was it solved upstream?
> > 
> 
> those two functions are moved to arch/x86/kernel/apic/io_apic.c
> 
> maybe you can put that patch to 2.6.31.xx?

Ah, yeah, that might be good, I'll see about that, if I do another
.31-stable release, which right now, I do not think so.

thanks,

greg k-h

Re: [2.6.31.13] Build failure at arch/x86/kernel/apic/io_apic.c

[Greg KH]

On Mon, May 03, 2010 at 10:50:41AM -0700, Yinghai wrote:
> On 05/03/2010 10:44 AM, Greg KH wrote:
> > On Mon, May 03, 2010 at 10:16:06AM -0700, Yinghai wrote:
> >> On 05/01/2010 11:07 AM, Greg KH wrote:
> >>> On Mon, Apr 12, 2010 at 01:10:59PM +0900, Tetsuo Handa wrote:
> >>>> Commit d539e5576605d048e6aeb21cbe3a8e71dc5eea81 "x86: Fix SCI on IOAPIC != 0"
> >>>> introduced "void setup_IO_APIC_irq_extra(u32 gsi)" which calls
> >>>> mp_find_ioapic() and mp_find_ioapic_pin().
> >>>>
> >>>> This commit does not compile if CONFIG_X86_IO_APIC=y and CONFIG_ACPI=n .
> >>>
> >>> Why would you want to build without ACPI on any modern system?  Anyway,
> >>> have a fix for this?
> >>>
> >> maybe just put #ifdef CONFIG_ACPI around the callee etc?
> > 
> > Ick.  How was it solved upstream?
> > 
> 
> those two functions are moved to arch/x86/kernel/apic/io_apic.c
> 
> maybe you can put that patch to 2.6.31.xx?

In looking at it further, that patch is just too big for .31 at this
point in time, sorry.  Just select ACPI as a config option, and you
should be fine.

Actually, use .32 or .33, I wouldn't recommend using .31 anymore anyway.

thanks,

greg k-h