Re: [PATCH] drivers/arch_numa: remove rebudant setup_per_cpu_areas()

[Pingfan Liu]

On Tue, Mar 9, 2021 at 10:02 PM Will Deacon <will@xxxxxxxxxx> wrote:
>
> [typo in subject "rebudant"]
>
> On Tue, Mar 09, 2021 at 06:21:38PM +0800, Pingfan Liu wrote:
> > There are two identical implementations of setup_per_cpu_areas() in
> > mm/percpu.c and drivers/base/arch_numa.c.
> >
> > Hence removing the one in arch_numa.c. And let arm64 drop
> > HAVE_SETUP_PER_CPU_AREA.
> >
> > Signed-off-by: Pingfan Liu <kernelfans@xxxxxxxxx>
> > Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
> > Cc: Will Deacon <will@xxxxxxxxxx>
> > Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > Cc: "Rafael J. Wysocki" <rafael@xxxxxxxxxx>
> > Cc: Atish Patra <atish.patra@xxxxxxx>
> > Cc: linux-kernel@xxxxxxxxxxxxxxx
> > To: linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
> > ---
> >  arch/arm64/Kconfig       |  4 ----
> >  drivers/base/arch_numa.c | 22 ----------------------
> >  2 files changed, 26 deletions(-)
> >
> > diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> > index 1f212b47a48a..d4bf8be0c3d5 100644
> > --- a/arch/arm64/Kconfig
> > +++ b/arch/arm64/Kconfig
> > @@ -1022,10 +1022,6 @@ config USE_PERCPU_NUMA_NODE_ID
> >       def_bool y
> >       depends on NUMA
> >
> > -config HAVE_SETUP_PER_CPU_AREA
> > -     def_bool y
> > -     depends on NUMA
> > -
> >  config NEED_PER_CPU_EMBED_FIRST_CHUNK
> >       def_bool y
> >       depends on NUMA
> > diff --git a/drivers/base/arch_numa.c b/drivers/base/arch_numa.c
> > index 4cc4e117727d..23e1e419a83d 100644
> > --- a/drivers/base/arch_numa.c
> > +++ b/drivers/base/arch_numa.c
> > @@ -167,28 +167,6 @@ static void __init pcpu_fc_free(void *ptr, size_t size)
> >  {
> >       memblock_free_early(__pa(ptr), size);
> >  }
> > -
> > -void __init setup_per_cpu_areas(void)
> > -{
> > -     unsigned long delta;
> > -     unsigned int cpu;
> > -     int rc;
> > -
> > -     /*
> > -      * Always reserve area for module percpu variables.  That's
> > -      * what the legacy allocator did.
> > -      */
> > -     rc = pcpu_embed_first_chunk(PERCPU_MODULE_RESERVE,
> > -                                 PERCPU_DYNAMIC_RESERVE, PAGE_SIZE,
> > -                                 pcpu_cpu_distance,
> > -                                 pcpu_fc_alloc, pcpu_fc_free);
>
> This doesn't look identical to the version in mm/percpu.c -- that one passes
> NULL instead of 'pcpu_cpu_distance' and tries to allocate the pcpu memory on
> the relevant NUMA nodes. In fact, if you could remove this function, you
> could probably remove the whole HAVE_SETUP_PER_CPU_AREA block here as the
> other functions are just used as helpers. So I'm not sure this is valid.
>
You are right. I need to rethink about it to see whether these two
functions can be unified into one.

Thanks,
Pingfan