Re: [PATCH] sparc/mm: export symbols for lazy_mmu_mode KUnit tests

[Andreas Larsson]

On 2026-01-12 11:09, Kevin Brodsky wrote:
> On 18/12/2025 11:05, Kevin Brodsky wrote:
>> Upcoming KUnit tests will call lazy_mmu_mode_{enable,disable}.
>> These tests may be built as a module, and because of inlining this
>> means that arch_{enter,flush,leave}_lazy_mmu_mode need to be
>> exported.
>>
>> Signed-off-by: Kevin Brodsky <kevin.brodsky@xxxxxxx>
>> ---
>>
>> The CI reports that sparc needs pretty much the same time treatment as
>> powerpc... Here's another patch to take care of that.
>>
>> Andrew, could you please add it after the powerpc one? At this point it's
>> probably best to remove the comment above MODULE_IMPORT_NS() in
>> mm/tests/lazy_mmu_mode_kunit.c. Thank you and sorry for the noise!
> 
> Gentle ping - I think we need this patch in mm-unstable, the CI has been
> complaining as well. Thanks!

I get this problem as well, which is solved by this patch.

>> ---
>>  arch/sparc/mm/tlb.c | 6 ++++++
>>  1 file changed, 6 insertions(+)
>>
>> diff --git a/arch/sparc/mm/tlb.c b/arch/sparc/mm/tlb.c
>> index 3a852071d260..6d9dd5eb1328 100644
>> --- a/arch/sparc/mm/tlb.c
>> +++ b/arch/sparc/mm/tlb.c
>> @@ -11,6 +11,8 @@
>>  #include <linux/preempt.h>
>>  #include <linux/pagemap.h>
>>  
>> +#include <kunit/visibility.h>
>> +
>>  #include <asm/tlbflush.h>
>>  #include <asm/cacheflush.h>
>>  #include <asm/mmu_context.h>
>> @@ -54,6 +56,8 @@ void arch_enter_lazy_mmu_mode(void)
>>  {
>>  	preempt_disable();
>>  }
>> +/* For lazy_mmu_mode KUnit tests */
>> +EXPORT_SYMBOL_IF_KUNIT(arch_enter_lazy_mmu_mode);
>>  
>>  void arch_flush_lazy_mmu_mode(void)
>>  {
>> @@ -62,12 +66,14 @@ void arch_flush_lazy_mmu_mode(void)
>>  	if (tb->tlb_nr)
>>  		flush_tlb_pending();
>>  }
>> +EXPORT_SYMBOL_IF_KUNIT(arch_flush_lazy_mmu_mode);
>>  
>>  void arch_leave_lazy_mmu_mode(void)
>>  {
>>  	arch_flush_lazy_mmu_mode();
>>  	preempt_enable();
>>  }
>> +EXPORT_SYMBOL_IF_KUNIT(arch_leave_lazy_mmu_mode);
>>  
>>  static void tlb_batch_add_one(struct mm_struct *mm, unsigned long vaddr,
>>  			      bool exec, unsigned int hugepage_shift)
>>
>> base-commit: 8f0b4cce4481fb22653697cced8d0d04027cb1e8
Regardless of if it becomes a patch of its own in the end, or is
folded into another:

Acked-by: Andreas Larsson <andreas@xxxxxxxxxxx>

Cheers,
Andreas