Hi Andrew,

Thanks for taking a look.

Sorry, the previous version was based on mm-everything.

I've now rebased onto mm-new and resent as v2.

I compared against
https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git/log/?h=mm-new
and the code matches the patch context.

v2:https://lore.kernel.org/all/20260812092856.55296-1-liuqiangneo@163.com/

Thanks

Best Regards

Qiang


At 2026-08-12 09:24:40, "Andrew Morton" <akpm@linux-foundation.org> wrote:
>On Tue, 11 Aug 2026 17:22:55 +0800 liuqiangneo@163.com wrote:
>
>> From: Qiang Liu <liuqiang@kylinos.cn>
>> 
>> Move pr_debug() inside the `if (dpage)` block to avoid a NULL deref,
>> and fix the direction label from "sys to dev" to "dev to sys" to match
>> the device-to-system copy.
>> 
>> ...
>>
>> --- a/lib/test_hmm.c
>> +++ b/lib/test_hmm.c
>> @@ -1151,10 +1151,9 @@ static vm_fault_t dmirror_devmem_fault_alloc_and_copy(struct migrate_vma *args,
>>  		if (!dpage && !order)
>>  			return VM_FAULT_OOM;
>>  
>> -		pr_debug("migrating from sys to dev pfn src: 0x%lx pfn dst: 0x%lx\n",
>> -				page_to_pfn(spage), page_to_pfn(dpage));
>> -
>>  		if (dpage) {
>> +			pr_debug("migrating from dev to sys pfn src: 0x%lx pfn dst: 0x%lx\n",
>> +				 page_to_pfn(spage), page_to_pfn(dpage));
>>  			lock_page(dpage);
>>  			*dst |= migrate_pfn(page_to_pfn(dpage));
>>  		}
>
>Current kernel code doesn't look like this?