Re: [PATCH 0/1] sched/topology: NUMA distance deduplication

Hi Sergei!

On 3/17/21 10:14 PM, Sergei Trofimovich wrote:
> Or is it just an early bootstrap message assuming more are to come?
> 
> Could it be that we initialize too little of generic acpi boilerplace
> when there is no SRAT? Somewhere around:
> 
>     https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/ia64/kernel/acpi.c#n446
> 
> arm64 and riscv calls `arch_numa_init()` and initializes numa node
> data in numa_init().
> While ia64 only calls acpi_numa_init() but not arch_acpi_numa_init()
> (which would do numa_init() for us).
> 
> x86 seems to vaguely resemble generic code by duplicating numa_init().
> 
> (UNTESTED) Could it be that we need something like:
> 
> --- a/arch/ia64/kernel/setup.c
> +++ b/arch/ia64/kernel/setup.c
> @@ -571,7 +571,7 @@ setup_arch (char **cmdline_p)
>         acpi_table_init();
>         early_acpi_boot_init();
>  #ifdef CONFIG_ACPI_NUMA
> -       acpi_numa_init();
> +       arch_numa_init();
>         acpi_numa_fixup();
>  #ifdef CONFIG_ACPI_HOTPLUG_CPU
>         prefill_possible_map();

That doesn't build:

  CALL    scripts/atomic/check-atomics.sh
  CALL    scripts/checksyscalls.sh
<stdin>:1511:2: warning: #warning syscall clone3 not implemented [-Wcpp]
  CHK     include/generated/compile.h
  CC      arch/ia64/kernel/setup.o
  CC      arch/ia64/kernel/acpi.o
arch/ia64/kernel/setup.c: In function 'setup_arch':
arch/ia64/kernel/setup.c:574:2: error: implicit declaration of function 'arch_numa_init'; did you mean 'acpi_numa_init'? [-Werror=implicit-function-declaration]
  574 |  arch_numa_init();
      |  ^~~~~~~~~~~~~~
      |  acpi_numa_init
cc1: some warnings being treated as errors
make[1]: *** [scripts/Makefile.build:271: arch/ia64/kernel/setup.o] Error 1
make[1]: *** Waiting for unfinished jobs....
make: *** [Makefile:1851: arch/ia64/kernel] Error 2