Re: [PATCH v2] m68k: Define NR_CPUS

[David Hildenbrand <david@redhat.com>]

On 24.09.24 09:48, Geert Uytterhoeven wrote:
> Hi David,
> 
> On Tue, Sep 24, 2024 at 9:34 AM David Hildenbrand <david@redhat.com> wrote:
>> On 24.09.24 01:56, Guenter Roeck wrote:
>>> v2: Instead of trying to make SPLIT_PTE_PTLOCKS depend on the
>>>       existence of NR_CPUS, define NR_CPUS for m68k.
>>
>> Okay, looks like we're cleaning up CONFIG_NR_CPUS for good.
>>
>> I'm back from conference travel tomorrow; I'll then throw in the following
>> into cross compilers and fixup any other arch that needs attention:
>>
>>
>> diff --git a/include/linux/threads.h b/include/linux/threads.h
>> index 1674a471b0b4..e31715e6746b 100644
>> --- a/include/linux/threads.h
>> +++ b/include/linux/threads.h
>> @@ -13,8 +13,7 @@
>>     * bit of memory.  Use nr_cpu_ids instead of this except for static bitmaps.
>>     */
>>    #ifndef CONFIG_NR_CPUS
>> -/* FIXME: This should be fixed in the arch's Kconfig */
>> -#define CONFIG_NR_CPUS 1
>> +#error "CONFIG_NR_CPUS not defined"
>>    #endif
>>
>>    /* Places which use this should consider cpumask_var_t. */
> 
> This is gonna trigger on almost all architectures if CONFIG_SMP=n.

Right, probably it's not as easy as:

diff --git a/mm/Kconfig b/mm/Kconfig
index 09aebca1cae3..b9566312fc9c 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -581,6 +581,13 @@ endif # MEMORY_HOTPLUG
  config ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE
         bool
  
+if !SMP
+
+config NR_CPUS
+       default 1
+
+endif # !SMP
+
  # Heavily threaded applications may benefit from splitting the mm-wide
  # page_table_lock, so that faults on different parts of the user address
  # space can be handled with less contention: split it at this NR_CPUS.


-- 
Cheers,

David / dhildenb