Re: [PATCH 6/8] irq: make io_apic_set_pci_routing to take device

[Ingo Molnar <mingo@elte.hu>]

* Yinghai Lu <yinghai@kernel.org> wrote:

> Impact: so use dev_to_node

> and use node in irq_desc_all_node()
> 
> Signed-off-by: Yinghai Lu <yinghai@kernel.org>
> 
> ---
>  arch/ia64/kernel/acpi.c        |    5 +++--
>  arch/x86/include/asm/io_apic.h |    4 ++--
>  arch/x86/include/asm/mpspec.h  |    4 +++-
>  arch/x86/kernel/acpi/boot.c    |    8 ++++----
>  arch/x86/kernel/apic/io_apic.c |   10 ++++++++--
>  drivers/acpi/pci_irq.c         |    5 +++--
>  drivers/char/hpet.c            |    4 ++--
>  drivers/pnp/pnpacpi/rsparser.c |    2 +-
>  include/linux/acpi.h           |    2 +-
>  9 files changed, 27 insertions(+), 17 deletions(-)

Please split this into two patches: one "obviously correct" one that 
just extends the call parameters, and another one that makes use of 
the new information:

> +	if (dev)
> +		node = dev_to_node(dev);
> +	else
> +		node = cpu_to_node(boot_cpu_id);

Btw., this sequence raises the question: wouldnt it be cleaner to 
have some sort of (somewhat artificial) 'boot platform device' that 
could be passed down instead of NULL, which would give 
dev_to_node(&boot_platform_dev) == boot_cpu_id?

Dunno - maybe it's not worth it and your solution is the cleanest.

Also, when submitting patches, please always try to Cc: affected 
non-x86 maintainers:

  scripts/get_maintainer.pl -f drivers/acpi/pci_irq.c   \
   drivers/pnp/pnpacpi/rsparser.c include/linux/acpi.h  \
   arch/ia64/kernel/acpi.c

Thanks,

	Ingo