Re: [PATCH] arm64, pci, numa: Adding helper functions as required by pci

[Bjorn Helgaas]

On Sat, Jan 09, 2016 at 11:35:17AM +0530, Ganapatrao Kulkarni wrote:
> On Fri, Jan 8, 2016 at 11:36 PM, Will Deacon <will.deacon@xxxxxxx> wrote:
> > On Fri, Jan 08, 2016 at 12:01:31PM -0600, Bjorn Helgaas wrote:
> >> On Tue, Dec 22, 2015 at 04:08:50PM +0530, Ganapatrao Kulkarni wrote:
> >> > Adding helper functions and necessary code to make
> >> > pci driver(pci-host-generic) numa aware.
> >> >
> >> > This patch is on top of arm64-numa v7.
> >> > http://www.spinics.net/lists/arm-kernel/msg460813.html
> >> >
> >> > Signed-off-by: Ganapatrao Kulkarni <gkulkarni@xxxxxxxxxxxxxxxxxx>
> >> > ---
> >> >  arch/arm64/include/asm/numa.h       | 8 +++++---
> >> >  arch/arm64/kernel/pci.c             | 8 ++++++++
> >> >  drivers/pci/host/pci-host-generic.c | 1 +
> >> >  3 files changed, 14 insertions(+), 3 deletions(-)
> >> >
> >> > diff --git a/arch/arm64/include/asm/numa.h b/arch/arm64/include/asm/numa.h
> >> > index b8c2a3f..1eca4db 100644
> >> > --- a/arch/arm64/include/asm/numa.h
> >> > +++ b/arch/arm64/include/asm/numa.h
> >> > @@ -15,9 +15,11 @@
> >> >  extern int __node_distance(int from, int to);
> >> >  #define node_distance(a, b) __node_distance(a, b)
> >> >
> >> > -/* dummy definitions for pci functions */
> >> > -#define pcibus_to_node(node)       0
> >> > -#define cpumask_of_pcibus(bus)     0
> >> > +struct pci_bus;
> >> > +extern int pcibus_to_node(struct pci_bus *bus);
> >> > +#define cpumask_of_pcibus(bus)     (pcibus_to_node(bus) == -1 ?            \
> >> > +                            cpu_all_mask :                         \
> >> > +                            cpumask_of_node(pcibus_to_node(bus)))
> >>
> >> All the arch definitions of cpumask_of_pcibus() are pretty similar;
> >> could they be made more generic?  At the very least, can you drop the
> >> definition here and use the one in include/asm-generic/topology.h,
> >> which is identical to what you're defining here?
> macro cpumask_of_pcibus is defined under #ifndef CONFIG_NUMA in
> include/asm-generic/topology.h
> and this macro is defined again for NUMA in respective arch header files.
> i think moving this macro from ifndef will allow us to use without redefinition.
> 
> --- a/include/asm-generic/topology.h
> +++ b/include/asm-generic/topology.h
> @@ -54,14 +54,14 @@
>  #define pcibus_to_node(bus)    ((void)(bus), -1)
>  #endif
> 
> +#endif /* CONFIG_NUMA */
> +
>  #ifndef cpumask_of_pcibus
>  #define cpumask_of_pcibus(bus) (pcibus_to_node(bus) == -1 ?            \
>                                  cpu_all_mask :                         \
>                                  cpumask_of_node(pcibus_to_node(bus)))
>  #endif
> 
> -#endif /* CONFIG_NUMA */
> -
>  #if !defined(CONFIG_NUMA) || !defined(CONFIG_HAVE_MEMORYLESS_NODES)
> 
>  #ifndef set_numa_mem
> 
> i have got below compilation error withour macro defined in arch code.
> 
> drivers/pci/pci-sysfs.c: In function âcpuaffinity_showâ:
> drivers/pci/pci-sysfs.c:114:15: error: implicit declaration of
> function âcpumask_of_pcibusâ [-Werror=implicit-function-declaration]
> drivers/pci/pci-sysfs.c:114:34: warning: initialization makes pointer
> from integer without a cast [enabled by default]
> drivers/pci/pci-sysfs.c: In function âcpulistaffinity_showâ:
> drivers/pci/pci-sysfs.c:123:34: warning: initialization makes pointer
> from integer without a cast [enabled by default]
> cc1: some warnings being treated as errors
> 
> 
> i can send this as separate patch( as a  fix patch).

Unless your series has already been merged, I don't like the idea of a
separate fix patch.  The whole point of reviews is to try to fix as
many things as we can *before* merging.

Bjorn