Re: [PATCH V3] selftests: vm: Add test for Soft-Dirty PTE bit

[Gabriel Krisman Bertazi]

Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx> writes:

> On 3/3/22 11:39 AM, Gabriel Krisman Bertazi wrote:
>> Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx> writes:
>> 
>>> On 2/28/22 2:37 AM, David Hildenbrand wrote:
>>>> On 24.02.22 22:23, Muhammad Usama Anjum wrote:
>>>>> This introduces three tests:
>>>>> 1) Sanity check soft dirty basic semantics: allocate area, clean, dirty,
>>>>> check if the SD bit flipped.
>>>>> 2) Check VMA reuse: validate the VM_SOFTDIRTY usage
>>>>> 3) Check soft-dirty on huge pages
>>>>>
>>>>> This was motivated by Will Deacon's fix commit 912efa17e512 ("mm: proc:
>>>>> Invalidate TLB after clearing soft-dirty page state"). I was tracking the
>>>>> same issue that he fixed, and this test would have caught it.
>>>>>
>>>> A note that madv_populate.c already contains some SOFTDIRTY tests
>>>> regarding MADV_POPULATE. Eventually we want to factor out
>>>> softdirty/pagemap handling+checks for easier reuse.
>>>>
>>>
>>> Is this patch unnecessary then?
>> It is not unnecessary since the madv test doesn't cover the bug tested
>> here, afaik.  But, as mentioned when I originally submitted this patch,
>> it should be merged into selftests/vm/madv_populate.c or, at least,
>> reuse that existing infrastructure.
>> https://lore.kernel.org/lkml/87lf553z5g.fsf@xxxxxxxxxxxxx/
>> 
>
> Oops this one came in a few months ago and appears to have slipped
> through and didn't get the right attention. Sorry about that.
>
> Please resend the patch and cc all the everybody on this thread.
>
> I would like to have your patch reviewed and looked at first. This
> patch needs rework sine it has several comments to be addressed.

Hi Shuah,

The patch being discussed in that thread is the same that Usama is
proposing here, minus a few modifications. Usama has taken over the work
to upstream it.

We just spoke, and he will follow up with a new version that addresses
the coding issues and reuses the infrastructure from madv_populate.c

Sorry for the noise.

-- 
Gabriel Krisman Bertazi